An approach to the precise chemoenzymatic synthesis of 13C-labeled sialyloligosaccharide on an intact glycoprotein: A novel one-pot [3-13C]-labeling method for sialic acid analogues by control of the reversible aldolase reaction, enzymatic synthesis of [3-13C]-NeuAc-α-(2→3)-[U- 13C]-Gal-β-(1-4)-GlcNAc-β-sequence onto glycoprotein

Article abstract of DOI:10.1021/ja994211j

A one-pot enzymatic ¹³C-labeling method for the 3-position of sialic acid (NeuAc) analogues has been developed using NeuAc aldolase, lactate dehydrogenase (LDH), alcohol dehydrogenase (ADH), and nucleotide pyrophosphatase (NPP). This method consists of two steps, the first of which is degradation to 2-acetamido-2-deoxy-D-mannose (ManNAc) analogues. This degradation reaction was accelerated by a cofactor regeneration system which converts pyruvic acid into lactic acid using LDH, ADH, and β-nicotinamide adenine dinucleotide oxidized form (β-NAD⁺). The second step is condensation of the ManNAc analogue with [3-¹³C]-pyruvic acid newly added after decomposition of the cofactor by nucleotide pyrophosphatase which play a role like switch to stop conversion of pyruvic acid into lactic acid. Five different NeuAc analogues have been labeled in good yields using this newly developed one-pot enzymatic procedure. Following conversion of [3-¹³C]-NeuAc to CMP-[3-¹³C]-NeuAc, enzymatic synthesis of [3-¹³C]-NeuAc-α-(2→3)-[U- ¹³C]-Gal-β-(1→4)-GlcNAc-β-x-ovalbumin (x: hybrid type oligosaccharide) 23 and [3-¹³C]-NeuAc-α-(2→3)-[U- ¹³C]-Gal-β-(1→4)-GlcNAc-β-OMe 26 (sialyl LacNAc) was performed using bovine β-1,4-galactosyltransferase and rat recombinant α-2,3-sialyltransferase. The ¹H chemical shifts of all protons in [3-¹³C]-NeuAc-α-(2→3)-[U-¹³C]-Gal-β-on a glycoprotein were assigned by 2D HMQC, 1D HSQC-TOCSY, and the herein described 1D and 2D HSQC-TOCSY-NOESY-TOCSY method. More specifically, the 7-, 8-, and 9-protons of NeuAc could be observed by this HSQC-TOCSY-NOESY-TOCSY method even with only a single ¹³C atom at the 3-position. In addition, 1D and 2D HMQC-NOESY spectra as well as carbon spin-lattice relaxation times (T₁) were measured to compare the conformational properties and dynamic behavior of the sialylgalactoside as part of the sialyl LacNAc 26 and when bound to a glycoprotein 23. These analyses suggested that the conformational properties of sialyl LacNAc are similar for both the conjugated and unconjugated forms, and that the torsional angle of the sialyl linkage, i.e., COOH-C2^NeuAc-O-C3^Gal, is biased toward the anti (-146.7°) conformation. In addition, the flexibility of galactosyl ring when bound to a glycoprotein appears to be significantly restricted by the attachment of NeuAc as compared with unconjugated sialyl LacNAc.

Full text of DOI:10.1021/ja994211j

5678
J. Am. Chem. Soc. 2000, 122, 5678-5694
An Approach to the Precise Chemoenzymatic Synthesis of
13C-Labeled Sialyloligosaccharide on an Intact Glycoprotein: A
Novel One-Pot [3-13C]-Labeling Method for Sialic Acid Analogues by
Control of the Reversible Aldolase Reaction, Enzymatic Synthesis of
[3-13C]-NeuAc-R-(2f3)-[U-13C]-Gal-â-(1f4)-GlcNAc-â- Sequence
onto Glycoprotein, and Its Conformational Analysis by Developed
NMR Techniques
Tatsuo Miyazaki,† Hajime Sato,‡ Tohru Sakakibara,† and Yasuhiro Kajihara*,†
Contribution from the Graduate School of Integrated Science, and Faculty of Science, Yokohama City
UniVersity, 22-2, Seto, Kanazawa-ku, Yokohama 236-0027 Japan, and Bruker Japan Co., Ltd. 21-5,
Ninomiya, 3-chome, Tsukuba, Ibaraki 305-0051 Japan
ReceiVed December 2, 1999
Abstract: A one-pot enzymatic 13C-labeling method for the 3-position of sialic acid (NeuAc) analogues has
been developed using NeuAc aldolase, lactate dehydrogenase (LDH), alcohol dehydrogenase (ADH), and
nucleotide pyrophosphatase (NPP). This method consists of two steps, the first of which is degradation to
2-acetamido-2-deoxy-D-mannose (ManNAc) analogues. This degradation reaction was accelerated by a cofactor
regeneration system which converts pyruvic acid into lactic acid using LDH, ADH, and â-nicotinamide adenine
dinucleotide oxidized form (â-NAD+). The second step is condensation of the ManNAc analogue with [3-13C]-
pyruvic acid newly added after decomposition of the cofactor by nucleotide pyrophosphatase which play a
role like switch to stop conversion of pyruvic acid into lactic acid. Five different NeuAc analogues have been
labeled in good yields using this newly developed one-pot enzymatic procedure. Following conversion of
[3-13C]-NeuAc to CMP-[3-13C]-NeuAc, enzymatic synthesis of [3-13C]-NeuAc-R-(2f3)-[U-13C]-Gal-â-(1f4)-
GlcNAc-â-x-ovalbumin (x: hybrid type oligosaccharide) 23 and [3-13C]-NeuAc-R-(2f3)-[U-13C]-Gal-â-(1f4)-
GlcNAc-â-OMe 26 (sialyl LacNAc) was performed using boVine â-1,4-galactosyltransferase and rat recombinant
R-2,3-sialyltransferase. The 1H chemical shifts of all protons in [3-13C]-NeuAc-R-(2f3)-[U-13C]-Gal-â- on a
glycoprotein were assigned by 2D HMQC, 1D HSQC-TOCSY, and the herein described 1D and 2D HSQC-
TOCSY-NOESY-TOCSY method. More specifically, the 7-, 8-, and 9-protons of NeuAc could be observed
by this HSQC-TOCSY-NOESY-TOCSY method even with only a single 13C atom at the 3-position. In addition,
1D and 2D HMQC-NOESY spectra as well as carbon spin-lattice relaxation times (T1) were measured to
compare the conformational properties and dynamic behavior of the sialylgalactoside as part of the sialyl
LacNAc 26 and when bound to a glycoprotein 23. These analyses suggested that the conformational properties
of sialyl LacNAc are similar for both the conjugated and unconjugated forms, and that the torsional angle of
the sialyl linkage, i.e., COOH-C2NeuAc-O-C3Gal, is biased toward the anti (-146.7°) conformation. In addition,
the flexibility of galactosyl ring when bound to a glycoprotein appears to be significantly restricted by the
attachment of NeuAc as compared with unconjugated sialyl LacNAc.
Introduction
electrostatic nature of the hydroxyl and carboxyl groups are
central to the molecular recognition events mediated by binding
of sialyloligosaccharide to receptor proteins. However, although
the conformational properties2b,l,p,r of sialyloligosaccharide
analogues of low molecular weight have been reported by many
research groups,2 the conformation and dynamics of sialyloli-
(1) (a) Yogeeswaran, G.; Salk, P. L. Science 1981, 212, 1514-1516.
(b) Wiley, D. C.; Skehel, J. J. Annu. ReV. Biochem. 1987, 56, 365-394.
(c) Rademacher, T. W.; Parekh, R. B.; Dwek, R. A. Annu. ReV. Biochem.
1988, 57, 785-838. (d) Polley, M. J.; Phillips, M. L.; Wayner, E.;
Nudelman, E.; Singhal, A. K.; Hakomori, S.; Paulson, J. C. Proc. Natl.
Acad. Sci. U.S.A. 1991, 88, 6224-6228. (e) Pilatte, Y.; Bignon, J.; Lambre´,
C. R. Glycobiology 1993, 3, 201-217. (f) Varki, A. Glycobiology 1993, 3,
97-130. (g) Springer, T. A. Cell 1994, 76, 301-314. (h) Varki, A. Proc.
Natl. Acad. Sci. U.S.A. 1994, 91, 7390-7397. (i) McEver, R. P.; Moore,
K. L.; Cummings, R. D. J. Biol. Chem. 1995, 270, 11025-11028. (j) Dwek,
R. A. Chem. ReV. 1996, 96, 683-720.
Sialic acid occurs frequently as a terminal residue of
carbohydrate chains on glycolipids and glycoproteins in cell
membranes. The sialic acid residues at the nonreducing end of
oligosaccharides are known to be associated with several
recognition events in carbohydrate-mediated cell adhesion.1
Therefore, research into the role of sialic acid is the focus of
many research groups around the world.2 For this purpose,
gangliosides, N- and O-linked type sialyloligosaccharides, and
their analogues have been synthesized,2c-e,h,m,s and their binding
ability during several molecular recognition events have been
assayed.2a,f,g,i-k,n,o,q Conformational properties as well as the
† Yokohama City University.
‡ Bruker Japan Co.
10.1021/ja994211j CCC: $19.00 © 2000 American Chemical Society
Published on Web 05/31/2000

Synthesis of Sialyloligosaccharide on Glycoprotein
J. Am. Chem. Soc., Vol. 122, No. 24, 2000 5679
gosaccharides and their analogues attached to glycoproteins have
not yet been fully analyzed.
gosaccharide analogues attached to glycoproteins using NMR,
just as if structures of oligosaccharides of low molecular weights
are analyzed. To address this problem, we developed a concise
[3-13C]-labeling method for NeuAc analogues, and reported
preliminary results6 for the synthesis of the 13C-labeled sialy-
loligosaccharide, 9-deoxy-9-fluoro-[3-13C]-NeuAc-R-(2f6)-[U-
13C]-Gal-â-(1f4)-GlcNAc-â- sequence on an intact glycopro-
tein using a CMP-[3-13C]-NeuAc analogue and UDP-[U-13C]-
Glc, and its analysis by NMR measurement.
The purpose of the research described herein is to develop a
synthetic technique for preparing sialyloligosaccharides and their
analogues on intact glycoproteins with precision and to analyze
their conformations, thus providing valuable new information
on the conformation and dynamic behavior of these sialyloli-
gosaccharides and their role in binding to receptor proteins.
Recently, enzymatic methods3 have been used successfully
to attach sugar analogues to the nonreducing end of oligosac-
charides on glycoproteins.4 The drawback to this approach is
the difficulty of analyzing short sugar analogue sequences of a
much larger glycoprotein. Moreover, structural analysis of the
nonreducing end of sugar residues on glycoproteins has not been
successful using X-ray crystallography,1j and it is difficult to
identify the sugar protons due to overlap with the resonances
of the amino acids using NMR.5 Therefore, it would be very
essential to develop a concise analytic method for sialyloli-
13C-labeled NeuAc7 has been utilized for conformational
analysis of sialyloligosaccharides on artificial membrane
surfaces7a,b and TRNOE experiments.7f These 13C-labeled
NeuAc compounds were synthesized by condensation of 13C-
labeled pyruvic acid with an excess of ManNAc.7a However,
to synthesize 13C-labeled NeuAc analogues, such as the 5-, 7-,
8-, and 9-modified NeuAc, corresponding 2-, 4-, 5-, and
6-modified ManNAc analogues must first be prepared. To date,
a large number of NeuAc analogues have been synthesized from
NeuAc,8 the common synthetic route including selective protec-
tion of hydroxyl groups has become fairly common, unlike the
synthesis of ManNAc analogues. Additionally, the substrate
specificity of NeuAc aldolase is flexible to a variety of NeuAc
and ManNAc analogues.3d,9 Using this knowledge, if [3-13C]-
NeuAc analogues can be obtained from NeuAc derivatives, these
established routes8 for synthesis of the NeuAc analogues can
be practically utilized. The reason the 3-position of NeuAc
should be 13C-labeled is that we hope to use novel NMR
techniques for observation of the H-3 to H-9 protons. In a
previous report, we were not able to observe the H-7 proton of
9-deoxy-9-fluoro-[3-13C]-NeuAc because HMQC-TOCSY de-
velopment stopped at H-6. Therefore, 13C-labeling of the side
chain of NeuAc was thought necessary. However, if a combined
(2) (a) Paulson, J. C.; Rogers, G. N.; Carroll, S. M.; Higa, H. H.; Pritchett,
T.; Milks, G.; Sabeasan, S. Pure Appl. Chem. 1984, 56, 797-805. (b) Weis,
W.; Brown, J. H.; Cusack, S.; Paulson, J. C.; Skehel, J. J.; Wiley, D. C.
Nature 1988, 333, 426-431. (c) Hasegawa, A.; Murase, T.; Adachi, K.;
Morita, M.; Ishida, H.; Kiso, M. J. Carbohydr. Chem. 1990, 9, 181-199.
(d) Hasegawa, A.; Adachi, K.; Yoshida, M.; Kiso, M. Carbohydr. Res. 1992,
230, 273-288. (e) Hasegawa, A.; Kiso, M. Carbohydrates Synthetic
Methods and applications in Medicinal Chemistry; Ogura, H., Hasegawa,
A., Suami, T., Eds.; Kodansha: Tokyo, 1992; pp 243-266. (f) Sabeasan,
S.; Duus, J. Ø.; Neira, S.; Domaille, P.; Kelm, S.; Paulson, J. C.; Bock, K.
J. Am. Chem. Soc. 1992, 114, 8363-8375. (g) von Itzstein, M.; Wu, W.-
Y.; Kok, G. B.; Pegg, M. S.; Dyason, J. C.; Jin, B.; Phan, T. V.; Smythe,
M. L.; White, H. F.; Oliver, S. W.; Colman, P. M.; Varghese, J. N.; Ryan,
D. M.; Woods, J. M.; Bethell, R. C.; Hotham, V. J.; Cameron, J. M.; Penn,
C. R. Nature 1993, 363, 418-423. (h) Fujita, S.; Numata, M.; Sugimoto,
M.; Tomita, K.; Ogawa, T. Carbohydr. Res. 1994, 263, 181-196. (i)
Sabesan, S.; Neira, S.; Davidson, F.; Duus, J. Ø. Bock, K. J. Am. Chem.
Soc. 1994, 116, 1616-1634. (j) DeFrees, S. A.; Kosch, W.; Way, W.;
Paulson, J. C.; Sabesan, S.; Halcomb, R. L.; Huang, D.-H.; Ichikawa, Y.;
Wong, C.-H. J. Am. Chem. Soc. 1995, 117, 66-79. (k) Sabesan, S.; Neira,
S.; Wasserman, Z. Carbohydr. Res. 1995, 267, 239-261. (l) Scheffer, K.;
Ernst, B.; Katopodis, A.; Magnani, J. L.; Wang, W. T.; Weisemann, R.;
Peters, T. Angew. Chem., Int. Ed. Engl. 1995, 34, 1841-1844. (m) Yoshida,
M.; Ishida, H.; Kiso, M.; Hasegawa, A. Carbohydr. Res. 1996, 280, 331-
338. (n) Sigal, G. B.; Mammen, M.; Dahmann, G.; Whitesides, G. M. J.
Am. Chem. Soc. 1996, 118, 3789-3800. (o) DeFrees, S. A.; Phillips, L.;
Guo, L.; Zalipsky, S. J. Am. Chem. Soc. 1996, 118, 6101-6104. (p) Ng,
K. K.-S.; Weis, W. I. Biochemistry 1997, 36, 979-988. (q) Kamitakahara,
H.; Suzuki, T.; Nishigori, N.; Suzuki, Y.; Kanie, O.; Wong, C.-H. Angew.
Chem., Int. Ed. 1998, 37, 1524-1528. (r) Henrichsen, D.; Ernst, B.;
Magnani, J. L.; Wang, W.-T.; Meyer, B.; Peters, T. Angew. Chem., Int.
Ed. 1999, 38, 98-102. (s) Komba, S.; Galustian, C.; Ishida, H.; Feizi, T.;
Kannagi, R.; Kiso, M. Angew. Chem., Int. Ed. 1999, 38, 1131-1133. (t)
Sears, P.; Wong, C.-H.; Angew. Chem., Int. Ed. 1999, 38, 2300-2324.
(3) (a) Hindsgaul, O.; Kaur, K. J.; Gokhale, U. B.; Srivastava, G.; Alton,
G.; Palcic, M. M. Enzymes in Carbohydrate Synthesis; Bednarski, M. D.,
Simon, E. S., Eds.; ACS Symposium Series 466; American Chemical
Society: Washington, DC, 1991; pp 38-50. (b) Kodama, H.; Kajihara,
Y.; Endo, T.; Hashimoto, H. Tetrahedron Lett. 1993, 34, 6419-6422. (c)
Brossmer, R.; Gross, H. J. Methods in Enzymology; Lee, Y. C., Lee, R. T.,
Eds.; Academic Press: New York, 1994; Vol. 247, pp 177-193. (d) Wong,
C.-H.; Whitesides, G. M. Enzymes in Synthetic Organic Chemistry; Baldwin,
J. E., Magnus, P. D., Eds.; Pergamon: New York, 1994.
(5) A few reseach groups have performed to assign protons of oligosac-
charide chain on an intact glycoprotein (<35 kDa) by using an advanced
NMR technique: (a) Brockbank, R. L.; Vogel, H. J.; Biochemistry 1990,
29, 5574-5583. (b) de Beer, T.; van Zuylen, C. W. E. M.; Ha˚rd, K.;
Boelens, R.; Kaptein, R.; Kamerling, J. P.; Vliegenthart, J. F. G. FEBS
Lett. 1994, 348, 1-6. (c) Wyss, D. F.; Choi, J. S.; Wagner, G. Biochemistry
1995, 34, 1622-1634. (d) Wyss, D. F.; Choi, J. S.; Li, J.; Knoppers, M.
H.; Willis, K. J.; Arulanandam, A. R. N.; Smolyar, A.; Reinherz, E. L.
Science 1995, 269, 1273-1278.
(6) Miyazaki, T.; Sakakibara, T.; Sato, H.; Kajihara, Y. J. Am. Chem.
Soc. 1999, 121, 1411-1412.
(7) (a) Aubin, Y.; Prestegard, J. H. Biochemistry 1993, 32, 3422-3428.
(b) Salvatore, B. A.; Ghose, R.; Prestegard, J. H. J. Am. Chem. Soc. 1996,
118, 4001-4008. (c) Gervay, J.; Mamuya, N. N.; Barber, R. A.; Tetrahedron
Lett. 1997, 38, 1865-1868. (d) Probert, M. A.; Milton, M. J.; Harris, R.;
Schenkman, S.; Brown, J. M.; Homans, S. W.; Field, R. A. Tetrahedron
Lett. 1997, 38, 5861-5864. (e) Tietze, L. F.; Gretzke, D. Eur. J. Org. Chem.
1998, 1895-1899. (f) Harris, R.; Kiddle, G. R.; Field, R. A.; Milton, M.
J.; Ernst, B.; Magnani, J. L.; Homans, S. W. J. Am. Chem. Soc. 1999, 121,
2546-2551.
(8) (a) Zbiral, E. Carbohydrates Synthetic Methods and Applications in
Medicinal Chemistry; Ogura, H., Hasegawa, A., Suami, T., Eds.; Kodan-
sha: Tokyo, 1992; pp 304-339. (b) von Itzstein, M.; Thomson, R. J. Top.
Curr. Chem. 1997, 186, 119-170.
(4) (a) Conradt, H. S.; Bu¨nsch, A.; Brossmer, R. FEBS Lett. 1984, 170,
295-300, (b) Higa, H. H.; Paulson, J. C. J. Biol. Chem. 1985, 260, 8838-
8849. (c) Berman, E.; Lis, H.; James, T. L. Eur. J. Biochem. 1986, 161,
589-594. (d) Berman, E.; Lis, H. Biochim. Biophys. Acta 1987, 924, 403-
407. (e) Gross, H. J.; Bu¨nsch, A.; Paulson, J. C.; Brossmer, R. Eur. J.
Biochem. 1987, 168, 595-602. (f) Gross, H. J.; Brossmer, R. Glycocon-
jugate J. 1987, 4, 145-156. (g) Gross, H. J.; Brossmer, R. Eur. J. Biochem.
1988, 177, 583-589. (h) Herrler, G.; Gross, H. J.; Imhof, A.; Brossmer,
R.; Milks, G.; Paulson, J. C. J. Biol. Chem. 1992, 267, 12501-12505. (i)
Gilhespy-Muskett, A. M.; Partridge, J.; Jefferis, R.; Homans, S. W.
Glycobiology 1994, 4, 485-489. (j) Kajihara, Y.; Endo, T.; Ogasawara,
H.; Kodama, H.; Hashimoto, H. Carbohydr. Res. 1995, 269, 273-294. (k)
Gross, H. J.; Brossmer, R. Glycoconjugate J. 1995, 12, 739-746. (l) Witte,
K.; Sears, P.; Martin, R.; Wong, C.-H. J. Am. Chem. Soc. 1997, 119, 2114-
2118. (m) Yamaguchi, Y.; Kato, K.; Shindo, M.; Aoki, S.; Furusho, K.;
Koga, K.; Takahashi, N.; Arata, Y.; Shimada, I. J. Biomol. NMR 1998, 12,
385-394.
(9) Reviews for synthesis of NeuAc analogues by aldolase: (a) Toone,
E. J.; Simon, E. S.; Bednarski, M. D.; Whitesides, G. M. Tetrahedron, 1989,
45, 5365-5422. (b) Wong, C.-H. Enzymes in Carbohydrate Synthesis;
Bednarski, M. D., Simon, E. S., Eds.; ACS Symposium Series 466;
American Chemical Society: Washington, DC, 1991; pp 23-37. (c)
Drueckhammer, D. G.; Hennen, W. J.; Pederson, R. L.; Barbas, C. F., III;
Gautheron, C. M.; Krach, T.; Wong, C.-H. Synthesis 1991, 499-525. (d)
Wong, C.-H.; Halcomb, R. L.; Ichikawa, Y.; Kajimoto, T. Angew. Chem.,
Int. Ed. Engl. 1995, 34, 412-432. (e) Auge´, C.; Narvor, C. G.-L.
PreparatiVe Carbohydrate Chemistry; Hanessian, S., Eds.; Dekker, New
York, 1996, 467-484. (f) Gijsen, H. J. M.; Qiao, L.; Fitz, W.; Wong, C.-
H. Chem. ReV. 1996, 96, 443-473. (g) Fessner, W.-D.; Walter, C. Top.
Curr. Chem. 1996, 184, 97-194. (h) Lin, C.-C, Lin, C.-H, Wong, C.-H.
Tetrahedron Lett. 1997, 38, 2649-2652. (i) Wu, W.-Y.; Jin, B.; Kong, D.
C. M.; von Itzstein, M. V. Carbohydr. Res. 1997, 300, 171-174. (j) Kuboki,
A.; Okazaki, H.; Sugai, T.; Ohta, H. Tetrahedron 1997, 53, 2387-2400.

5680 J. Am. Chem. Soc., Vol. 122, No. 24, 2000
Miyazaki et al.
Scheme 1a
a Reagents: (a) (1) Tf2O, pyridine, CH2Cl2; (2) TASF, CH2Cl2, y ) 52% (2 steps); (b) (1) phenyl chlorothionoformate, pyridine, CH2Cl2; (2)
AIBN, n-Bu3SnH, toluene, 80 °C, y ) 39% (2 steps); (c) (1) 10% Pd-C, AcOH; (2) 0.3 N NaOH: MeOH ) 1:1; (3) 25 mM HCl, Amberlyst 15
(H+) 80 °C, y ) 64% (3 steps); (d) (1) NBS, acetone: H2O ) 9:1; (2) 0.3 N NaOH, y ) 52% (2 steps); (e) (1) Ba(OH)2‚8H2O, BaO, BnBr, DMF;
(2) CH2N2; (3) 60% AcOH, 60 °C, y ) 62% (3 steps); (f) AcCl, pyridine: CH2Cl2 ) 1:1, y ) 82%; (g) (1) phenyl chlorothionoformate, DMAP,
pyridine: CH2Cl2 ) 1:1; (2) AIBN, n-Bu3SnH, toluene, 100 °C, y ) 68% (2 steps); (h) (1) 10% Pd-C, AcOH; (2) 0.3 N NaOH: MeOH ) 2:1; (3)
25 mM HCl, Amberlyst 15 (H+) 85 °C, y ) 59% (3 steps); (i) BzCl, pyridine: CH2Cl2 ) 1:4, y ) 68%; (j) (1) 1,1′-thiocarbonyldiimidazole,
DMAP, CH2Cl2; (2) AIBN, n-Bu3SnH, toluene, 80 °C, y ) 80% (2 steps); (k) (1) 60% AcOH, 60 °C; (2) 0.3N NaOH: MeOH ) 1:1; (3) 25 mM
HCl, Amberlyst 15 (H+) 80 °C, y ) 69% (3 steps).
TOCSY-NOESY-TOSCY10 and HSQC technique, that is,
HSQC-TOCSY-NOESY-TOCSY, could be developed, all pro-
tons of the NeuAc from H-3 to H-9 would be measurable even
with only a single 13C-atom at 3-position of NeuAc which bound
to a glycoprotein. During the synthesis of NeuAc analogues by
multistep sequences, 13C-labeling of several carbons of the sugar
skeleton would require expensive, time-consuming, advanced
techniques. Therefore, the development of such a novel mea-
surement technique promised to enable us to simplify the
synthetic routes for 13C-labeling.
In this paper, we would like to describe a novel one-pot
[3-13C]-labeling method for NeuAc analogues, the chemoen-
zymatic synthesis of the [3-13C]-NeuAc-R-(2f3)-[U-13C]-Gal-
â- sequence on an intact glycoprotein, and its conformational
analysis with a newly developed HSQC-TOCSY-NOESY-
TOCSY technique which requires only one 13C-labeled carbon
at the 3-position of NeuAc.
pyruvic acid should be incorporated into the first three carbons
(C1-C2-C3) of the NeuAc analogue since the NeuAc aldolase
reaction is reversible. As a simple 13C-labeling method for
NeuAc analogues, we first examined this exchange reaction
using the following condition: NeuAc analogues (15 µmol), 3
equiv of [3-13C]-pyruvic acid (45 µmol) and 10 U of NeuAc
aldolase (Figure 1). The degree of 13C-labeling was determined
by measurement of the integration values for the both the H-3
proton of [3-13C]-NeuAc analogue (Figure 2A) and the CH3-
1
protons of [3-13C]-pyruvic acid (Figure 2B) by H NMR. The
7- and 8-deoxy-[3-13C]-NeuAc analogues could not be estimated
by comparison of the integration values due to overlap between
the H-3NeuAc proton with other protons. Therefore, the degree
of mixing between [3-13C]-pyruvic acid and normal pyruvic acid
was monitored, and the results are shown in Figure 2B. NeuAc,
9-deoxy-9-fluoro (9-F), 9-azido-9-deoxy (9-N3), and 9-deoxy-
NeuAc analogues quickly achieved equilibrium and were labeled
at rates of roughly 75% each (Figure 2A). In contrast, exchange
with the 7- and 8-deoxy derivatives proceeded very slowly
(Figure 2B). These analogues were labeled at lower rates even
after 90 h incubation. After purification on an anion-exchange
column, the labeling rate of all NeuAc analogues were
determined, and the results are shown in Table 1. The labeling
rates ranged from 62 to 77% for the 8-, and 9-modified NeuAc
compounds. In contrast, the 7-deoxy-analogue showed only 8%
Results and Discussion
Chemical Synthesis of NeuAc Analogues. For this study,
we prepared five different NeuAc analogues (9-deoxy-9-fluoro-
NeuAc 3, 9-deoxy-NeuAc 5, 9-deoxy-9-azido-NeuAc 8, 8-deoxy-
NeuAc 13, 7-deoxy-NeuAc 16) as shown in Scheme 1. This
scheme represents slightly modified routes of several previously
reported synthetic routes for NeuAc analogues.11
(11) For synthesis of 9-F-NeuAc: (a) Sharma, M.; Petrie, C. R., III;
Korytnyk, W. Carbohydr. Res. 1988, 175, 25-34. For synthesis of deoxy-
NeuAc: (b) Brandstetter, H. H.; Zbiral, E. Liebigs Ann. Chem. 1983, 12,
2055-2274. (c) Hasegawa, A.; Adachi, K.; Yoshida, M.; Kiso, M.
Carbohydr. Res. 1992, 230, 257-272. For synthesis of 9-N3-NeuAc: (d)
Brossmer, R.; Rose, U. Biochem. Biophys. Res. Commun. 1980, 96, 1282-
1289. (e) Kajihara, Y.; Ebata, T.; Kodama. H. Angew. Chem., Int. Ed. 1998,
37, 3166-3169.
Simple Labeling Method Consisting of Exchange Reaction
between a NeuAc Analogue and a [3-13C]-Pyruvic Acid
Mediated by NeuAc Aldolase. When aldolase is added to a
mixture of a NeuAc analogue and [3-13C]-pyruvic acid, [3-13C]-
(10) Uhr´ın, D.; Brisson, J.-R.; Kogan, G.; Jennings, H. J. Magn. Res.
1994, 104, 289-293.

Synthesis of Sialyloligosaccharide on Glycoprotein
J. Am. Chem. Soc., Vol. 122, No. 24, 2000 5681
Figure 1. [3-13C]-labeling method for NeuAc analogues by an
exchange reaction using NeuAc aldolase.
enrichment. To determine why the labeling rates ranged from
8 to 77%, we measured the Km and Vmax values for the aldolase
degradation of synthetic NeuAc analogues.12 As shown in Table
1, NeuAc and the 9-modified analogues which had good labeling
rates have small Km values (2.9-4.6 mM) and higher velocities.
In contrast, the 7-deoxy derivative had a large Km value and
lower velocity. Therefore, the labeling rate appears to depend
on the Vmax/Km values of the NeuAc derivatives. The addition
of more enzyme might have improved the labeling rates but
would have been impractical for larger scale reactions. There-
fore, we concluded that this exchange reaction was not suitable
for use as a general [3-13C]-labeling method for NeuAc
analogues but was useful for the synthesis of 14C-labeled NeuAc
analogues.
Acceleration of Degradation Reaction by NeuAc Aldolase.
The problem with the [3-13C]-labeling method for NeuAc
analogues is that reaction velocities of some analogues are very
slow. We then tried to accelerate the aldolase-mediated degrada-
tion step. The velocity of the aldolase reaction depends on the
quantity of pyruvic acid generated. Therefore, we added lactate
dehydrogenase (LDH) and 1 equiv of â-nicotinamide adenine
dinucleotide, reduced form (â-NADH), to remove the pyruvic
Figure 2. Degree of 13C-labeling in the 3-position of NeuAc analoguesa
a
and (A) pyruvic acid (B) during exchange reaction. 7- and 8-Deoxy-
NeuAc are not contained. bPopulations were estimated using integration
1
of H-3 proton in H NMR spectra.
Table 1. Kinetic Parameters for NeuAc Aldolase Reacting with
Several NeuAc Analogues in Degradation Reactions
13C-labeling
(%)
Km
Vmax
substrate
NeuAc
9-F-NeuAc
9-deoxy-NeuAc
9-N3-NeuAc
8-deoxy-NeuAc 13
7-deoxy-NeuAc 16
(mM) (µmol/min) Vmax/Km
74
75
75
77
62
8
4.4
4.6
3.0
2.9
19
28
56.0
42.3
41.5
43.3
10.4
4.8
12.7
9.2
13.8
14.9
0.55
0.17
3
5
8
acid, as lactic acid, from the equilibrium mixture. Preliminary
results using this approach have been reported for three different
analogues, the 9-F-, 9-N3-, and 8-deoxy-analogues.6 By this
strategy, following degradation of NeuAc to ManNAc, the
ManNAc derivatives were isolated and then condensed with
[3-13C]-pyruvic acid. Although the velocity of degradation was
indeed accelerated under these conditions, the 8-deoxy-NeuAc
was degraded to 5-deoxy-ManNAc in 45% yield. Further
investigation revealed that the 7- and 8-deoxy NeuAc analogues
were smoothly degraded to the corresponding ManNAc ana-
logues after 9 h of incubation with 1.5 equiv of â-NADH. These
reactions were monitored by measurement of the remaining
NeuAc in the reaction mixture using the thiobarbituric acid
(12) Although the kinetic parameters of these NeuAc analogues were
reported, Km and Vmax values of 9-F, 9-N3, 9-deoxy, 8-deoxy, and 7-deoxy
NeuAc toward NeuAc aldolase (E. coli) had not been measured under the
same conditions: (a) Gantt, R.; Millner, S.; Binkley, S. B. Biochemistry
1964, 3, 1952-1960. (b) Schauer, R.; Stoll, S.; Zbiral, E.; Schreiner, E.;
Brandstetter, H.; Vasella, A.; Baumberger, F. Glycoconjugate J. 1987, 4,
361-369. (c) Aisaka, K.; Igarashi, A.; Yamaguchi, K.; Uwajima, T.
Biochem. J. 1991, 276, 541-546. (d) Zbiral, E.; Kleineidam, R. G.;
Schreiner, E.; Hartmann, M.; Christian, R.; Schauer, R. Biochem. J. 1992,
282, 511-516.

5682 J. Am. Chem. Soc., Vol. 122, No. 24, 2000
Miyazaki et al.
quantitatively degraded to ManNAc using the â-NADH regen-
eration system (â-NAD+ of 0.01 equiv based on NeuAc). After
1.5 h, the reaction velocity was found to be roughly equal to
that when 1.5 equiv of â-NADH is used. We then showed that
NPP quantitatively hydrolyzes pyrophosphate bonds. The hy-
drolysis reaction of the cofactors can be easily monitored by
TLC (solvent system: ethyl acetate/MeOH/H2O ) 3/2/1.5; Rf:
â-NAD+ ) 0.20, â-NADH ) 0.45).
On the basis of these results, we attempted a synthesis of
[3-13C]-NeuAc analogues by a one-pot enzymatic procedure,
and the results are summarized in Figure 6. The reactions were
monitored by measuring the quantity of NeuAc analogue by
TBA. The degradation reactions were run in solutions containing
30 mg of NeuAc analogue, 30 U of NeuAc aldolase, 150 U of
LDH and ADH, 120 µL of EtOH, and 0.01 equiv of â-NAD+
(based on NeuAc). After 20 h, the degradation reaction of the
NeuAc analogues using the cofactor regeneration system was
virtually complete. Then NPP (2.8 U) was added to the reaction
mixture to inactivate the LDH. After hydrolysis of pyrophos-
phates in â-NADH and â-NAD+, [3-13C]-pyruvic acid was
added. The recondensation reaction proceeded immediately and
appeared to achieve an equilibrium state after 20 h. After
purification of [3-13C]-NeuAc analogues by anion-exchange and
gel permeation column chromatography, the isolation yields and
13C-enrichment (%) were determined (Table 2). The yields of
the condensation reaction ranged from 55 to 76% and the degree
of 13C-labeling ranged from 87 to 97%. In particular, the [3-13C]-
7-deoxy-derivative was 91% labeled, dramatically higher than
that resulting from the exchange reaction (8%). This procedure
provides the desired [3-13C]-labeled NeuAc analogues with an
excellent degree of 13C-enrichment. In the condensation reaction,
the yield for formation of [3-13C]-NeuAc analogues estimated
by TBA (Figure 6), also depended on the Vmax/Km values. The
NeuAc, 9-deoxy-, 9-N3-NeuAc having the largest Vmax/Km value
achieved an equilibrium state in which the concentration of the
NeuAc analogue ranged from 10 to 15 mM in the mixture, as
shown in Figure 6. In contrast, the 9-F, 8-deoxy, and 7-deoxy
analogues having small Vmax/Km values achieved equilibrium
states in which the concentrations of NeuAc ranged from 15 to
20 mM. These results suggest that good substrates for degrada-
tion result in low condensation yields, but poor substrates result
in good condensation yields. Although in this experiment we
did not use 5-modified NeuAc, this method should also be useful
for the synthesis of 5-modified-[3-13C]-NeuAc derivatives.9h-j
Since this route allows for the preparation of the desired [3-13C]-
NeuAc analogues in moderate yields with high degrees of 13C-
labeling, we were able to proceed with the synthesis of 13C-
labeled sialyloligosaccharide containing glycoproteins.
Figure 3. Degradation of NeuAc analogues by use of NeuAc aldolase
and LDH (1.5 equiv of â-NADH).
assay13 (TBA) as shown in Figure 3. These results suggest that
the low yield for 8-deoxy-NeuAc under our previous reaction
conditions might have arisen from decomposition of â-NADH
due to the long reaction time. The accelerated velocity achieved
by use of LDH corresponds to the Vmax/Km values of NeuAc
analogues. Unfortunately, this method also proved to be
impractical for 13C-labeling of NeuAc analogues, because the
â-NADH complicates the purification of the ManNAc ana-
logues.
13C-Labeling Method for NeuAc Analogues by One-Pot
Enzymatic Procedure. The equilibrium bias for the aldolase
reaction is determined by the quantity of pyruvic acid in reaction
mixture. Therefore, we have developed a novel one-pot labeling
method which controls the quantity of pyruvic acid in the
reaction mixture as shown in Figure 4. The first step is
degradation of the NeuAc analogues accelerated by cofactor
regeneration.3d,14 The second step is recondensation of [3-13C]-
pyruvic acid with a ManNAc analogue following inactivation
of LDH. The key to this reaction is the use of nucleotide
pyrophosphatase (NPP) to remove â-NADH and the oxidized
form of â-nicotinamide adenine dinucleotide (â-NAD+). The
NPP reaction serves as a switch to stop conversion of pyruvic
acid into lactic acid. After decomposition of the cofactor,
addition of [3-13C]-pyruvic acid triggers [3-13C]-NeuAc forma-
tion. The merits of this strategy are the use of catalytic amounts
of â-NADH and the ability to perform this reaction as a one-
pot procedure. To determine whether this strategy was viable
or not, two points had to be confirmed: (1) how much â-NADH
in the cofactor regeneration is needed for degradation of NeuAc,
and (2) whether NPP effectively inactivates LDH. To regenerate
â-NADH from â-NAD+, we utilized ethanol and alcohol
dehydrogenase (ADH) because ethanol and the resulting ac-
etaldehyde are easily removed during the workup.
Synthesis of the [3-13C]-NeuAc-r-(2f3)-[U-13C]-Gal-â-x-
Ovalbumin 23 and [3-13C]-NeuAc-r-(2f3)-[U-13C]-Gal-â-
(1f4)-GlcNAc-â-OMe 26. Many examples of recognition
events mediated by R-(2f3) sialoside are now known such as
those involving SLeX, GM1, and GM4, and many conformational
studies of these ligands have been reported.15 Nevertheless the
differences in conformational properties and dynamic behavior
between sialosides bound to proteins and unconjugated sialy-
loligosaccharides is not yet well understood. Therefore, in this
study we wanted to synthesize a [3-13C]-NeuAc-R-(2f3)-[U-
13C]-Gal-â- sequence at the nonreducing end of a glycoprotein
(Scheme 2).
We first examined degradation of NeuAc by a cofactor
regeneration system. As shown in Figure 5, NeuAc was
(13) (a) Warren, L. J. Biol. Chem. 1959, 234, 1971-1975. (b) Aminoff,
D. Biochem. J. 1961, 81, 384-391.
(14) (a) Mosbach, K.; Larsson, P.-O.; Lowe, C. Methods Enzymol. 1976,
44, 859-887. (b) Shaked, Z.; Whitesides, G. M. J. Am. Chem. Soc. 1980,
102, 7104-7105. (c) Wong, C.-H.; Whitesides, G. M. J. Am. Chem. Soc.
1981, 103, 4890-4899. (d) Matos, J. R.; Wong, C.-H. J. Org. Chem. 1986,
51, 2388-2389. (e) Ichikawa, Y.; Shen, G.-J.; Wong, C.-H. J. Am. Chem.
Soc. 1991, 113, 4698-4700. (f) Ichikawa, Y.; Liu, J. L.-C.; Shen, G.-J.;
Wong, C.-H. J. Am. Chem. Soc. 1991, 113, 6300-6302. (g) Fang, J.-M.;
Wong, C.-H. Synlett 1994, 393-402.
Esteration and selective acetylation16 of [3-13C]-NeuAc 17
afforded derivative 18, which was then converted into cytidine
5′-monophospho-[3′′-13C]-NeuAc (CMP-[3′′-13C]-NeuAc) 20 by

Synthesis of Sialyloligosaccharide on Glycoprotein
J. Am. Chem. Soc., Vol. 122, No. 24, 2000 5683
Figure 4. Novel strategy for 13C-labeling method of NeuAc analogues by a one-pot enzymatic procedure.
examined using ovalbumin 21. It is known that boVine â-1,4-
galactosyltransferase also catalyzes the transfer of glucose from
UDP-Glc.19 Therefore, the use of UDP-[U-13C]-Glc risks
formation of a mixture of [U-13C]-Glc-â-(1f4)-GlcNAc and
[U-13C]-Gal-â-(1f4)-GlcNAc at the nonreducing end of the
glycoprotein. However, in our first trial to transfer the [U-13C]-
Gal residue from UDP-[U-13C]-Glc following treatment with
UDP-Glc-4-epimerase, no glucosyl linkages were observed in
the 2D HMQC spectrum.6 Furthermore, to obtain a NOE
between H-8NeuAc and H-3Gal, which is necessary for evaluation
of the population of the syn conformation of the sialyl linkage,15c
13C-labeling of the C-3Gal is essential for the HMQC-NOESY
measurement. However [U-13C]-Gal is much more difficult to
obtain than [U-13C]-Glc. Therefore, we used UDP-[U-13C]-Glc.
Since the ovalbumin has only one oligosaccharide chain
containing either one or two GlcNAc residues at the nonreducing
terminal,20 this glycoprotein was a good model substrate for
our research. Transfer of the [U-13C]-Gal residue proceeded
smoothly toward this glycoprotein.6 However, when the NeuAc
was transferred using the rat recombinant R-2,3-sialyltrans-
ferase, only 50% of the galactoside was estimated to be
sialylated according to the 2D HMQC spectrum despite repeated
sialylation reactions on this glycoprotein. Therefore, the [U-13C]-
Gal which was not sialylated was hydrolyzed by Diplococcus
pneumoniae â-galactosidase. Consequently the sialoside 23 was
obtained in an analytically pure state. We also prepared [3-13C]-
NeuAc-R-(2f3)-[U-13C]-Gal-â-(1f4)-GlcNAc-â-OMe 26, to
compare its conformational properties and dynamic behavior
with those of the glycoprotein.
Figure 5. Comparison of degradation of NeuAc analogues using
cofactor regeneration and with 1.5 equiv of â-NADH.
a method previously reported by this laboratory.17 To synthesize
the [3-13C]-NeuAc-R-(2f3)-[U-13C]-Gal-â- sequence on a
glycoprotein, enzymatic transfer of [U-13C]-Gal and [3-13C]-
NeuAc from the corresponding uridine 5′-diphospho-[U-13C]-
Glc (UDP-[U-13C]-Glc)18 and CMP-[3′′-13C]-NeuAc 20 was
(15) (a) Sabesan, S.; Bock, K.; Lemieux, R. U. Can. J. Chem. 1984, 62,
1034-1045. (b) Breg. J.; Kroon-batenburg, L. M. J.; Strecker, G.; Montreuil,
J.; Vliegenthart, J. F. G. Eur. J. Biochem. 1989, 178, 727-739. (c) Poppe,
L.; Dabrowski, J.; von der Lieth, C.-W.; Numata, M.; Ogawa, T. Eur. J.
Biochem. 1989, 180, 337-342. (d) Acquotti, D.; Poppe, L.; Dabrowski, J.;
von der Lieth, C.-W.; Sonnino, S.; Tettamanti, G. J. Am. Chem. Soc. 1990,
112, 7772-7778. (e) Scarsdale, J. N.; Prestegard, J. H.; Yu, R. K.
Biochemistry 1990, 29, 9843-9855, (f) Sabesan, S.; Bock, K.; Paulson, J.
C. Carbohydr. Res. 1991, 218, 27-54. (g) Sabesan, S.; Duus, J. Ø.;
Fukunaga, T.; Bock, K.; Ludvigsen, S. J. Am. Chem. Soc. 1991, 113, 3236-
3246. (h) Ichikawa, Y.; Lin, Y.-C.; Dumas, D. P.; Shen, G.-J.; Garcia-
Junceda, E.; Williams, M. A.; Bayer, R.; Ketcham, C.; Walker, L. E.;
Paulson, J. C.; Wong, C.-H. J. Am. Chem. Soc. 1992, 114, 9283-9298. (i)
Rutherford, T. J.; Spackman, D. G.; Simpson, P. J.; Homans, S. W.
Glycobiology 1994, 4, 59-68. (j) Cooke, R. M.; Hale, R. S.; Lister, S. G.;
Shah, G.; Weir, M. P. Biochemistry 1994, 33, 10591-10596. (k) Bernardi,
A.; Raimondi, L. J. Org. Chem. 1995, 60, 3370-3377. (l) Poppe, L.; Brown,
G. S.; Philo, J. S.; Nikrad, P. V.; Shah, B. H. J. Am. Chem. Soc. 1997, 119,
1727-1736. (m) Wu, W.-G.; Pasternack, L.; Huang, D.-H.; Koeller, K.
M.; Lin, C.-C.; Seitz, O.; Wong, C.-H. J. Am. Chem. Soc. 1999, 121, 2409-
2417.
1
Assignments of the H and 13C Chemical Shifts of the
[3-13C]-NeuAc-r-(2f3)-[U-13C]-Gal-â- on an Ovalbumin. A
2D HMQC experiment on [3-13C]-NeuAc-R-(2f3)-[U-13C]-
Gal-â-x-ovalbumin 23 (x: hybrid type) was performed to
analyze the 1H and 13C chemical shifts of the 13C-labeled
positions. Slices from the galactosyl part of the 2D HMQC
spectrum are shown in Figure 7 d-h. The differences in 13C
chemical shift values between galactoside 22 and sialoside 23
were measured in order to determine the sialylated position.
The chemical shift of the C-3Gal of galactoside 22 was shifted
downfield (2.93 ppm), and those for both C-2Gal and C-4Gal of
(16) Kuhn, R.; Lutz, P.; MacDonald, D. L. Chem. Ber. 1966, 99, 611-
617.
(17) Kajihara, Y.; Ebata, T.; Koseki, K.; Kodama, H.; Matsushita, H.;
Hashimoto, H. J. Org. Chem. 1995, 60, 5732-5735.
(18) (a) Goux, W. J.; Perry, C.; James, T. L. J. Biol. Chem. 1982, 257,
1829-1835. (b) Berman, E.; James, T. L. Carbohydr. Res. 1983, 113, 141-
150.
(19) Palcic, M. M.; Hindsgaul, O. Glycobiology 1991, 1, 205-209.
(20) (a) Huang, C.-C.; Mayer, H. E., Jr.; Montgomery, R. Carbohydr.
Res. 1970, 13, 127-137. (b) Tai, T.; Yamashita, K.; Ogata-Arakawa, M.;
Koide, N.; Muramatsu, T.; Iwashita, S.; Inoue, Y.; Kobata, A. J. Biol. Chem.
1975, 250, 8569-8575. (c) Tomiya, N.; Awaya, J.; Kurono, M.; Endo, S.;
Arata, Y.; Takahashi, N. Anal. Biochem. 1988, 171, 73-90.

5684 J. Am. Chem. Soc., Vol. 122, No. 24, 2000
Miyazaki et al.
Figure 6. Quantity (mM) of NeuAc analogues in the reaction mixture during a one-pot [3-13C]-labeling method.
Table 2. Results for the 13C-labeling Method of NeuAc Analogues
several NOESY mixing times (1, 50, 100, 200, 300, 400, and
500 ms). As shown in Figure 8, the NOE between H-6NeuAc
and H-7NeuAc clearly increased, depending on the mixing time,
and the maximum NOE as an inphase magnetization was
observed at τm ) 300 ms. By this measurement, the H-7NeuAc
chemical shift and coupling constant JH7,H8 were determined
(Figure 7b). Finally, we conducted the 1D HSQC-TOCSY-
NOESY-TOCSY experiment. Unfortunately, the H-6 and H-7
peaks overlapped, hindering our ability to selectively excite H-7
in the second TOCSY. As a result the 1D spectrum contained
both the desired broad H-7, -8, -9, and -9′ peaks as well as the
undesired H-4, -5, and -6 peaks, as shown in Figure 7c. In this
measurement, decreasing peak intensity results from two fac-
tors: (1) transverse magnetization is used in three of the
measurements (i.e., HSQC-TOCSY, TOCSY-NOESY, and
NOESY-TOCSY), and (2) selective excitation pulses22 are used
twice. As a result the intensity of the magnetization decreased
dramatically. In addition, when the magnetization of H-7 was
extracted, peak intensity of its magnetization had decreased to
less than 10% of the absolute intensity of H-3 as observed by
HSQC. Although we tried to optimize this measurement in order
to eliminate the undesired H-4, -5, and -6 peaks, we could not
obtain the desired spectrum. However the spectral pattern in
the 1D HSQC-TOCSY-NOESY-TOCSY spectrum obtained
here was almost identical to that reported for NeuAc in a
capsular polysaccharide.10 Despite this difficulty, assignment
of the H-8, -9, and -9′ peaks was roughly accomplished. We
then developed a 2D version of the HSQC-TOCSY-NOESY-
TOCSY spectrum (Figure 9). In this spectrum the resonance of
these cross-peaks could not be improved despite trying several
conditions. The desired cross-peaks for H-8, -9, and -9′ were
observed clearly in the expected chemical shift ranges and are
in good agreement with those of unconjugated sialyl LacNAc
26. The merit of this 2D measurement is that selective excitation
of the H-7 proton is not essential for the last TOCSY pulse
sequence. Therefore, the loss of transverse magnetization of H-7
is small compared to the 1D-version. In addition, TOCSY
development from H-7 to H-9, -9′ can be easily observed as a
well-separated cross-peak from TOCSY development of H-4,
by a One-Pot Enzymatic Procedure
isolated
conversion yield (%)a
degradation condens.
99 (68)c
98 (48)
99
99 (60)
95 (26)
96
yields 13C-labeling
substrate
NeuAc
9-F-NeuAc
9-deoxy-NeuAc
9-N3-NeuAc
8-deoxy-NeuAc 13
7-deoxy-NeuAc 16
(%)
(%)b
55
75
56
68
75
76
52
62
51
54
76
46
94 [74]d
96 [75]
96 [75]
97 [77]
87 [62]
91 [8]
3
5
8
a Conversion yields were determined by thiobarbituric acid assay
or HPLC. b The yields were determined by 1H NMR spectra of reaction
mixture. c The yields designated in parentheses are degradation yield
by use of only NeuAc aldolase.6 d The yields designated in brackets
are 13C-labeling yields for the exchange reaction using NeuAc aldolase.
galactoside 22 were shifted upfield (1.55 ppm, 1.05 ppm,
respectively) in the 13C NMR spectrum. The H-3Gal peak of
galactoside 22 in the 1H NMR spectrum was also shifted
downfield (0.47 ppm). Therefore, the sialylated position was
determined to be the 3-position of the galactoside 22.21 To assign
1
the H chemical shifts of other protons on the NeuAc residue,
a 1D HSQC-TOCSY experiment with selective excitation of
C-3NeuAc was performed, and the spectrum obtained is shown
Figure 7a (carbon decoupling was not used during acquisition).
Although the H-4NeuAc peak is broad, the 1H chemical shifts of
H-4NeuAc, H-5NeuAc, and H-6NeuAc, and the coupling constants
(JH4,H5, JH5,H6) were determined. As shown in Figure 7a, the
small coupling constant (JH6, H7 < ∼1.5 Hz) normally found in
NeuAc often makes it impossible to observe the correlations
between H-6NeuAc and H-7NeuAc in a TOCSY spectrum. There-
fore, it is also difficult to assign the H-7, -8, -9, and -9′ protons
with one 13C atom at the 3-position of a conjugated sialyloli-
gosaccharide. To overcome the problem, we developed a new
measurement technique. Homonuclear TOCSY-NOESY-TOC-
SY has recently been reported.10 This measurement technique
was used for NeuAc residues of a large molecular weight
capsular polysaccharide in order to observe all of the protons
of the NeuAc residues. The mechanism of this measurement
technique is transfer of H-6 magnetization, detected by the first
TOCSY from H-3NeuAc, to H-7 by NOESY and then the
magnetization of H-7 is transferred to H-9 by the second
TOCSY. If a combined HSQC and TOCSY-NOESY-TOCSY
experiment, that is, HSQC-TOCSY-NOESY-TOCSY, could be
developed, all protons from H-3 to H-9 of a NeuAc residue
could be assigned using one 13C atom at the 3-position of the
NeuAc residue, even on an intact glycoprotein. Therefore, we
conducted a 1D HSQC-TOCSY-NOESY measurement using
1
-5, and -6. The H- and 13C-chemical shifts and the coupling
constants for [U-13C]-Gal-â-x-ovalbumin 22, [3-13C]-NeuAc-
R-(2f3)-[U-13C]-Gal-â-x-ovalbumin 23, [U-13C]-Gal-â-(1f4)-
GlcNAc-â-OMe (LacNAc) 25, and [3-13C]-NeuAc-R-(2f3)-
[U-13C]-Gal-â-(1f4)-GlcNAc-â-OMe (sialyl LacNAc) 26 are
summarized in Table 3.
13C Spin-Lattice Relaxation Times of the [3-13C]-NeuAc-
r-(2f3)-[U-13C]-Gal-â-x-Ovalbumin 23 and [3-13C]-NeuAc-
(22) For selective pulses: (a) Kessler, H.; Oschkinat, H.; Griesinger, C.;
Bermel, W. J. Magn. Reson. 1986, 70, 106-133. (b) Kessler, H.; Mronga,
S.; Gemmecker, G. Magn. Reson. Chem. 1991, 29, 527-557.
(21) Bradbury, J. H.; Jenkins, G. A. Carbohydr. Res. 1984, 126, 125-
156.

Synthesis of Sialyloligosaccharide on Glycoprotein
J. Am. Chem. Soc., Vol. 122, No. 24, 2000 5685
Scheme 2a
a Reagents: (a) (1) Dowex 50W-X8 (H+), MeOH; (2) HClO4, Ac2O, y ) 37% (2 steps); (b) (1) 1H-tetrazole, MeCN, 19; (2) t-BuOOH, MeCN;
(3) DBU; (4) NaOMe, MeOH: H2O ) 1:2, y ) 35% (4 steps); (c) UDP-[U-13C]-glucose, UDP-glucose-4-epimerase, boVine â-1,4-galactosyltransferase;
(d) (1) rat recombinant R-2,3-(N)-sialyltransferase, 20; (2) D. pneumoniae â-galactosidase; (e) UDP-[U-13C]-glucose, UDP-glucose-4-epimerase,
boVine â-1,4-galactosyltransferase, y ) 57%; (f) rat recombinant R-2,3-(N)-sialyltransferase, 20, y ) 30%.
r-(2f3)-[U-13C]-Gal-â-(1f4)-GlcNAc-â-OMe 26. 13C spin-
lattice relaxation times (T1) have often been used to evaluate
the molecular dynamics of oligosaccharides.2i,23 The 13C atom
is a suitable nucleus for this purpose since the relaxation time
dramatically decreased for 240, 390, and 280 ms, respectively.
This phenomena indicates that the flexibility of the galactosyl
ring is severely restricted by the attachment of sialic acid to
the 3-position as compared with that of the unconjugated sialyl
LacNAc 26. The T1DD’s of the C-3NeuAc are similar for both
sialoside 23 and 26, and this suggests the flexibility of the
NeuAc ring is not affected by the attachment of the glycoprotein.
(T1DD) is dominated by carbon-proton dipolar relaxation for
DD
protonated carbons, and since NT1
and τc are inversely
proportional (where N is the number of protons attached to the
carbon).24 Therefore, we measured the T1
and T1 of the
obsd
DD
Conformational Properties of the [3-13C]-NeuAc-r-(2f3)-
[U-13C]-Gal-â-Ovalbumin 23 and Sialoside 26. To analyze
the conformational properties, we measured NOE by both 1D
and 2D HMQC-NOESY techniques. In the 2D HMQC-NOESY,
we observed NOEs to H-4NeuAc, H-5NeuAc, and H-3Gal from either
H-3axNeuAc or H-3eqNeuAc. We could not determine whether the
NOE was derived from H-3axNeuAc or H-3eqNeuAc, because the
2D HMQC-NOESY shows NOEs to the H-4NeuAc, H-5NeuAc, and
H-3Gal at the same chemical shifts as the C-3NeuAc in the F1
dimension (Figure 10f). Therefore, we modified the 1D HMQC-
NOESY in order to determine whether the NOE resulted from
H-3axNeuAc or H-3eqNeuAc. We changed the two rectangular
pulses (90°) for both proton and carbon to Gaussian and half-
Gaussian shaped selective pulses, respectively. The pulse
sequence is summarized in the Experimental Section (Figure
13). As shown in Figures 10b and 10d, we have 1D HMQC
spectra showing the individual excitations of H-3axNeuAc and
H-3eqNeuAc, respectively. We applied this magnetization to the
NOESY pulse sequence. The 1D HMQC-NOESY spectra show
that NOEs to H-3eqNeuAc, H-4NeuAc, H-5NeuAc, and H-3Gal resulted
from H-3axNeuAc (Figure 10c). NOEs from H-3eqNeuAc were
observed for H-3axNeuAc and H-4NeuAc (Figure 10e). These NOE
patterns are similar to those of unconjugated sialyl LacNAc 26
except for the NOE between H-3Gal and H-8NeuAc (Table 5,
H-3Gal excitation). In addition we also observed NOEs from
H-1Gal on the protein (Figure 7i). Since chemical shifts of these
13C labeled position in [U-13C]-Gal-â-x-ovalbumin 22, [3-13C]-
NeuAc-R-(2f3)-[U-13C]-Gal-â-x-ovalbumin 23, [U-13C]-Gal-
â-(1f4)-GlcNAc-â-OMe 25, and [3-13C]-NeuAc-R-(2f3)-[U-
13C]-Gal-â-(1f4)-GlcNAc-â-OMe 26 by the use of 100.61
MHz NMR for 13C nuclei (9.4 T). The T1 values for [U-13C]-
Gal-â-x-ovalbumin 22 were previously reported using a 25 MHz
spectrometer.18a However we conducted this measurement again
because the T1 value is dependent on the Larmor frequency.
obsd
All T1
values are averages of experiments performed in
triplicate and are summarized in Table 4. In the comparison
between [U-13C]-Gal-â-x-ovalbumin 22 and unconjugated Lac-
NAc 25, the difference of each T1DD value is small. Attachment
DD
of sialic acid to galactoside 25 caused a decrease in the T1
value for 230 ms at the 3-position of the galactoside, but the
DD
T1 of other carbons were only slightly decreased. However,
in the case of [3-13C]-NeuAc-R-(2f3)-[U-13C]-Gal-â-x-oval-
DD
bumin 23, the T1
values of C-2Gal, C-3Gal, and C-4Gal
(23) (a) Czarniecki, M. F.; Thornton, E. R. J. Am. Chem. Soc. 1977, 99,
8279-8282. (b) Berman, E.; Allerhand, A. J. Biol. Chem. 1980, 255, 4407-
4410. (c) Rutherford, T. J.; Partridge, J.; Weller, C. T.; Homans, S. W.
Biochemistry 1993, 32, 12715-12724. (d) Poppe, L.; van Halbeek, H.;
Acquotti, D.; Sonnino, S. Biophys. J. 1994, 66, 1642-1652. (e) van Zuylen,
C. W. E. M. T.; de Beer, T.; Leeflang, B. R.; Boelens, R.; Kaptein, R.;
Kamerling, J. P.; Vliegenthart, J. F. G. Biochemistry 1998, 37, 1933-1940.
(24) (a) Breitmaier, E.; Spohn, K.-H.; Berger, S. Angew. Chem., Int. Ed.
1975, 14, 144-159. (b) Czarniecki, M. F.; Thornton, E. R. J. Am. Chem.
Soc. 1977, 99, 8273-8279.

5686 J. Am. Chem. Soc., Vol. 122, No. 24, 2000
Miyazaki et al.
Figure 7. 1D NMR spectra of [3-13C]-NeuAc-R-(2f3)-[U-13C]-Gal-
â-x-ovalbumin 23. (a) 1D HSQC-TOCSY obtained with the pulse
sequence from Figure 13C. (b) 1D HSQC-TOCSY-NOESY (NOESY
mixing time: 300 ms) obtained with the pulse sequence from Figure
13D. (c) 1D HSQC-TOCSY-NOESY-TOCSY (NOESY mixing time
300 ms) obtained with the pulse sequence from Figure 13E. (d) A slice
of C-1Gal in 2D HMQC spectrum. (e) A slice of C-2Gal in 2D HMQC
spectrum. (f) A slice of C-3Gal and C-5Gal in 2D HMQC spectrum. (g)
A slice of C-4Gal in 2D HMQC spectrum. (h) A slice of C-6Gal in 2D
HMQC spectrum. (i) A slice of C-1Gal in 2D HMQC-NOESY spectrum
(NOESY mixing time: 100 ms). N: NeuAc, G: Gal, GN: GlcNAc.
The span of all 1D spectra (a-i) is identical.
Figure 8. 1D HSQC-TOCSY-NOESY spectra of sialoside 23 with
different mixing times.
sialyl linkages which have been found exist in a variety of syn:
anti ratios. For example, NeuAc-R-(2f3)-Gal-R,â-OH,15c NeuAc-
R-(2f3)-Gal-â-(1f4)-GlcNAc-â-Asn,15b and NeuAc-R-(2f3)-
Gal-â-(1f4)-GlcNAc-â-(CH2)8COOMe,15f exist in syn:anti
populations of 1:1, 3:1, and 0:1, respectively. Therefore, the
conformation of the R-(2f3)-sialyl linkage appears to have no
fixed standard conformation, and to be heavily dependent on
the aglycons at the reducing end of the galactoside and GlcNAc.
It is also reported that the NOE between H-3Gal and H-8NeuAc
,
and between H-3Gal and H-3axNeuAc are caused by syn (i.e.,
COOH-C2-O-C3 = -85°) and anti (i.e., COOH-C2-O-
C3 = -153°) conformers of the sialyl linkage, respectively.15c
Therefore, these NOEs are essential to an evaluation of syn:
anti populations. For NeuAc-R-(2f3)-Gal-â-(1f4)-GlcNAc-
â-OMe 26, we observed two typical NOEs, between H-3Gal and
H-8NeuAc and between H-3Gal and H-3axNeuAc (Table 5). However
the intensity of the NOE between H-3Gal and H-8NeuAc was found
to be very small in comparison with that between H-3Gal and
NOEs do not overlap with protons of the NeuAc residue or the
galactoside but are identical to the chemical shifts of H-4GlcNAc
,
H-6GlcNAc, and H-6′GlcNAc of 26, we assigned the observed NOEs
from H-1Gal are to be H-4GlcNAc, H-6GlcNAc, and H-6′GlcNAc. All
observed NOEs and build up rate curves for H-3axNeuAc of
conjugated form 23 and unconjugated 26 are summarized in
Table 5 and Figure 11, respectively.
Conformational properties of the common NeuAc-R-(2f3)-
Gal-â- sequence in oligosaccharides have been reported by
several research groups.15 In GM1, the sialyl linkage adopts a
near anti conformation due to steric hindrance of neighboring
branched disaccharide chain, Gal-â-(1f3)-GalNAc.15d Other
H-3axNeuAc. The ratio of these NOE volumes (H-3Gal/H-8NeuAc
:
H-3Gal/H-3axNeuAc) in 26 was estimated to be <1:8 by the use
of a build-up rate curve (NOESY mixing time: 100 ms, data

Synthesis of Sialyloligosaccharide on Glycoprotein
J. Am. Chem. Soc., Vol. 122, No. 24, 2000 5687
the distance between H-3Gal and H-3axNeuAc in the conjugated
form is found to be lengthy compared with those for the sialyl
linkages which adopt the anti conformation.15f There are two
possible reasons for this. In the first case the population of anti
conformer is slightly decreased due to equilibrium between syn
and anti conformations. In the latter case the major conformation
is the anti conformation in which distance between H-3Gal and
H-3axNeuAc is about 2.7 Å. However, since we could not prove
this, we simulated an average 3D structure of sialyl LacNAc in
which the distance between H-3Gal and H-3axNeuAc is constrained
to 2.7 Å. In addition, since the NOEs to H-4GlcNAc, H-6GlcNAc
,
and H-6′GlcNAc from H-1Gal are observed, the distance between
proton pairs (H-6GlcNAc, H-1Gal) and (H-6′GlcNAc, H-1Gal) can be
estimated to be 3.0 and 3.0 Å, respectively, and these values
were also used as constraints for simulation. The global
minimum structure (COOH-C2NeuAc-O-C3Gal ) -146.7°,
C2NeuAc-O-C3Gal-H3Gal ) -25.6°, H1Gal-C1Gal-O-C4GlcNAc
) 51.3°, C1Gal-O-C4GlcNAc-H4GlcNAc ) -7.4°), as shown in
Figure 12, was in good agreement with those of the previously
reported unconjugated sialyl LacNAc.15d,f
These results suggest that the conformational properties of
sialyloligosaccharides are not influenced by attachment to the
glycoprotein, but that the flexibility of the sugar backbone is
decreased. In our case, since the sialyloligosaccharide may be
exposed on protein surface, the conformational properties are
not very different from those of the unconjugated form. We
were able to obtain insight into the conformational properties
on a large molecular glycoprotein (>40 kDa) for the first time,
and we believed this methodology may be applied to different
conjugated sialyloligosaccharide-protein combinations.
Conclusions
In summary, we have described a concise 13C-labeling method
for 5-, 7-, 8-, and 9-modified NeuAc analogues using a one-
pot enzymatic procedure. This procedure is based on biasing
the equilibrium in the reversible aldolase reaction using cofactor
regeneration and reaction with a nucleotide pyrophosphatase.
This procedure successfully afforded five different [3-13C]-
NeuAc analogues in which the degree of 13C-labeling is 87%
or greater. This strategy enabled us to synthesize a [3-13C]-
NeuAc-R-(2f3)-[U-13C]-Gal-â- sequence by the use of glyco-
syltransferases on an intact glycoprotein. We also analyzed the
conformational properties of sialoside 23 on a glycoprotein using
several standard NMR techniques and developed a new proce-
dure, 1D HSQC-TOCSY-NOESY-TOCSY, for this purpose.
The 13C-labeled sialoside 23 bound to a glycoprotein gave us
highly useful information, namely, that the major conformation
of sialyl linkage is anti. We were able to demonstrate that a
combined chemical, enzymatic, and NMR technique allows for
the precise synthesis of a sialylgalactoside on an intact glyco-
protein. In addition, with only minimal 13C-enrichment (the
3-position of NeuAc residue and a [U-13C]-galactoside), we were
able to analyze the conformation of such glycoprotein-bound
sialylgalactosides, as if we had been dealing with a much smaller
molecular weight sialyloligosaccharide.
Figure 9. 2D HSQC-TOCSY-NOESY-TOCSY spectrum of sialoside
23. (a) 1D HSQC-TOCSY-NOESY (Figure 8e). (b) 1D HSQC-TOCSY-
NOESY-TOCSY (Figure 7c). Dashed line in the 2D spectrum shows
diagnal peaks.
are not shown). Therefore, these results suggest that conforma-
tion of unconjugated sialyl LacNAc 26 is biased to the anti
conformation. On the other hand, an NOE between H-3Gal and
H-8NeuAc was not observed in the conjugated sialylgalactoside
23 under various experimental conditions (for example, using
shorter or longer mixing times or increasing the scan time), but
an NOE between H-3Gal and H-3axNeuAc was clearly observed.
These results suggest that the sialyl linkage of the glycoprotein
also adopts a near anti conformation. The distance between
H-3axNeuAc and H-3Gal was then estimated according to a
reported procedure.25 As shown in Figure 11, since both the
NOE build up curves for H-3axNeuAc excited (A: conjugated,
B: unconjugated) are not saturated until at 50 ms of mixing
time, peak volume of H-3eqNeuAc and H-3Gal at 50 ms were used
Experimental Section
NMR spectra were recorded with JEOL EX-270 or Bruker AVANCE
400 instruments. The chemical shifts of 1H NMR are presented in ppm
and referenced to tetramethylsilane (δ ) 0.00 ppm) in CDCl3, HOD
(δ ) 4.69 ppm) in D2O, and HOD (δ ) 4.69 ppm) in CD3OD as an
internal standard. The chemical shifts of 13C NMR spectra are expressed
in ppm and referenced to CDCl3 (δ ) 77.00 ppm) in CDCl3, 1,4-dioxane
(δ ) 67.17 ppm, external) in D2O, and CD3OD (δ ) 49.80 ppm) in
CD3OD. Thin-layer chromatography (TLC) was used DC-Platten
to estimate the length (standard NOE H-3axNeuAc/H-3eqNeuAc
:
1.77 Å). The estimated distance between H-3Gal and H-3axNeuAc
was 2.7 Å (conjugated) and 3.0 Å (unconjugated). However,
(25) (a) Poppe, L.; von der Lieth, C.-H.; Dabrowski, J. J. Am. Chem.
Soc. 1990, 112, 7762-7771.

5688 J. Am. Chem. Soc., Vol. 122, No. 24, 2000
Table 3. 1H and 13C NMR Data of 22, 23, 25, and 26
Miyazaki et al.
residue
δa
(Ja)
22
25
23
26
R-D-NeuAc
H-3ax
H-3eq
H-4
H-5
H-6
H-7
H-8
H-9
H-9′
(3JH3ax,H3eq
)
1.78 (10.8)
2.75 (10.8)
3.67
3.83 (10.0)
3.62 (10.0)
3.58 (<1.5)b
3.90-3.83 (8.4)
3.90-3.83
3.63 (11.3)
1.77 (12.4)
2.73 (12.4)
3.67 (4.6)
3.82 (9.9)
3.62 (10.4)
3.57 (1.4)
3.86 (8.7)
3.85 (2.5)
3.62 (11.8)
(5.9)
(3JH3ax,H4
)
)
(2JH3eq,H4
(3JH4,H5
(3JH5,H6
(3JH6,H7
(3JH7,H8
(2JH8,H9
)
)
)
)
)
(2JH9′,H9
(3JH8,H9′
)
)
NAc
C-3
2.01
40.04 (130.2)
(134.9)
(1JC3,H3ax
(1JC3,H3eq
)
)
40.09 (128.4)
(134.3)
C-4
C-5
68.27
52.20
C-6
C-7
73.13
68.37
C-8
C-9
72.11
63.05
NAc
H-1
H-2
22.58
4.52 (7.8)
3.54
â-D-Gal
(3JH1,H2
)
4.43
3.52
4.44 (7.8)
3.51
4.53
3.56
H-3
H-4
3.63
3.89
3.64
3.90
4.10
3.95
4.08
3.93
H-5
H-6,6′
C-1
C-2
C-3
C-4
C-5
C-6
H-1
H-2
H-3
H-4
H-5
3.70
3.70
3.70
3.69
3.80-3.66
103.33 (45.5)
71.33 (39.3)
72.95 (38.5)
68.95 (38.5)
75.76 (44.5)
61.42
3.77-3.67
103.33 (45.9)
71.37 (39.7)
72.98 (38.7)
68.96 (38.6)
75.78 (44.5)
61.44
4.43 (7.6)
3.69
3.76-3.65
3.67
3.55 (2.2)
3.98 (5.2)
3.81 (12.3)
2.01
3.48
102.31
3.81-3.67
103.04 (47.9)
69.78 (40.3)
75.88 (38.4)
67.90 (38.4)
75.56 (44.7)
61.43
3.73-3.67
103.01 (47.1)
69.77 (40.3)
75.90 (38.4)
67.90 (38.4)
75.57 (44.3)
61.43
4.43 (7.8)
3.76-3.65
3.76-3.65
3.67
3.58 (2.1)
3.99 (5.7)
3.82 (12.1)
2.01
3.48
102.80
(1JC1,C2
(1JC2,C3
(1JC3,C4
(1JC4,C4
(1JC5,C6
)
)
)
)
)
â-D-GlcNAc
(3JH1,H2
)
3.69
(3JH5,H6
)
)
)
H-6
(3JH5,H6′
(2JH6,H6′
4.03
3.85
H-6′
H-NAc
H-OMe
C-1
C-2
C-3
55.42
73.02
55.40
73.07
C-4
C-5
79.09
75.46
78.58
75.41
C-6
C-OMe
C-NAc
60.63
57.53
22.61
60.65
57.52
22.58
1
a The chemical shifts and coupling constants were presented in ppm and Hz, respectively. H chemical shifts were measured at 303 K (internal
standard: HOD ) 4.69 ppm), 13C chemical shifts were measured at 303 K (external standard: 1,4-dioxane ) 67.17 ppm). b The HSQC-TOCSY
did not show a H-7NeuAc resonance from H-6NeuAc
.
Table 4. Relaxation Parameters T1obsd, NOE, and T1 for 13C Nuclei in D2O Solution at 303 Ka
DD
22
25
23
26
obsd
DD
obsd
DD
obsd
DD
obsd
DD
T1
(s)
NOE
(1+η)
T1
(s)
T1
(s)
NOE
(1+η)
T1
(s)
T1
(s)
NOE
(1+η)
T1
(s)
T1
(s)
NOE
(1+η)
T1
(s)
NeuAc
Gal
C-3
C-1
C-2
C-3
C-4
C-5
C-6
-
-
-
-
-
-
0.15
0.28
0.28
0.20
0.25
0.28
0.20
0.25
2.1
1.8
2.1
2.1
2.1
2.0
2.1
2.0
0.27
0.70
0.51
0.36
0.45
0.56
0.36
0.49
0.18
0.32
0.31
0.26
0.28
0.32
0.24
0.29
2.3
2.0
1.9
2.0
1.8
2.1
2.4
2.0
0.28
0.64
0.68
0.52
0.70
0.58
0.34
0.58
0.34
0.34
0.34
0.33
0.32
0.22
0.32
1.9
1.9
1.9
1.9
2.0
2.3
2.0
0.75
0.75
0.75
0.73
0.64
0.34
0.66
0.54
0.53
0.53
0.44
0.53
0.36
0.49
2.4
2.5
2.4
2.4
2.4
2.6
2.5
0.77
0.70
0.75
0.62
0.75
0.45
0.67
ave.
a T1 (s) ) (1.988/η)T1
(s).
DD
obsd
Kieselgel 60 F254 (Merck). Column chromatography was carried out
on Merck Silica gel 60 of 230-400 mesh.
Materials. The solvents and reagents were purified according to
standard procedures. N-acetyl-neuraminic acid aldolase (NeuAc aldo-
lase, EC 4.1.3.3) was obtained from TOYOBO Co., Ltd. Lactate
dehydrogenase (LDH, EC 1.1.1.27) and â-nicotinamide adenine di-
nucleotide oxidized form (â-NAD+) were from Oriental Yeast Co., Ltd.
[3-13C]-Sodium pyruvate (99.1% 13C-labeled) was obtained from

Synthesis of Sialyloligosaccharide on Glycoprotein
J. Am. Chem. Soc., Vol. 122, No. 24, 2000 5689
Figure 11. NOE build up curves caused by excitation of H-3axNeuAc
in conjugated 23 and unconjugated 26 by several mixing times.
Figure 10. 1D HMQC-NOESY spectrum of sialoside 23 with shaped
pulses for both H-3axNeuAc (or H-3eqNeuAc) and C-3NeuAc. (a) Standard
1D HMQC spectrum6 with selective excitation for C-3NeuAc. (b) 1D
HMQC (selective excitation for H-3axNeuAc) using modified pulse
sequence from Figure 13A. (c) 1D HMQC-NOESY (selective excitation
for H-3axNeuAc) using modified pulse sequence from Figure 13B. (d)
1D HMQC (selective excitation for H-3eqNeuAc) using the modified pulse
sequence from Figure 13A. (e) 1D HMQC-NOESY (selective excitation
for H-3eqNeuAc) using modified pulse sequence from Figure 13B. (f)
Slice of C-3NeuAc resonance in 2D HMQC-NOESY spectrum.
1
Table 5. Intra- and Inter-residual H/1H NOE in Sialoside 23 and
26
Figure 12. The global minimum energy conformation of conjugated
sialoside 23.
excitation
NeuAc H-3ax
23
NeuAc-H4, 5, Gal-H3 NeuAc-H4, 5, Gal-H3
NeuAc-H4 NeuAc-H4
Gal-H3, 5, GlcNAc-H4 Gal-H3, 5, GlcNAc-H4
26
were from Boehriger-Mannheim. Albumin, chicken egg (ovalbumin),
alcohol dehydrogenase (ADH, EC 1.1.1.1), nucleotide pyrophosphatase
(NPP, EC 3.6.1.9), â-nicotinamide adenine dinucleotide reduced form
(â-NADH), uridine 5′-diphosphoglucose 4-epimerase (EC 5.1.3.2), and
boVine â-1,4-galactosyltransferase (EC 2.4.1.22) were purchased from
Sigma. All other biochemical and chemical reagents were obtained from
Sigma and Aldrich, unless otherwise indicated.
Methyl (Methyl 5-Acetamido-4,7,8-tri-O-benzyl-3,5,9-trideoxy-
9-fluoro-r,â-D-glycero-D-galacto-2-nonulopyranosid) Onate (2).
9-Deoxy-9-fluoro-NeuAc 3 was synthesized by reported method11a
except for fluorination. To a cold (-10 °C) solution of compound 111a
(R:â ) 1:24, 1.46 g, 2.4 mmol) in CH2Cl2 (26.7 mL) and pyridine (0.6
mL) was added trifluoromethanesulfonic anhydride (1.0 mL, 5.9 mmol)
dropwise, and then the mixture was allowed to warm to room
H-3eq
H-1
Gal
GlcNAc-H6, 6′
Gal-H1
Gal-H1, 4
Gal-H3, 6
Gal-H3
GlcNAc-H6,6′
Gal-H1
Gal-H1, 4, NeuAc H8
Gal-H3, 6
H-2
H3
H-4
H-5
H-6/6′
Gal-H3
Gal-H4
Gal-H4
ISOTEC INC. Rat recombinant R-2,3-(N)-sialyltransferase (EC 2.4.99.5)
was from CALBIOCHEM. Calf intestinal alkaline phosphotase (CIAP,
EC 3.1.3.1) and Diplococcus pneumoniae â-galactosidase (EC 3.2.1.23)

5690 J. Am. Chem. Soc., Vol. 122, No. 24, 2000
Miyazaki et al.
temperature. After 45 min, the mixture was diluted with CH2Cl2 and
washed sequentially with 0.2 M HCl and water. The organic phase
was dried with MgSO4 and then concentrated in vacuo. The residue
was dried under vacuum (15 Pa) for 1 h. To a solution of this residue
in CH2Cl2 (12.6 mL) was added tris(dimethylamino)sulfonium difluo-
rotrimethylsilicate (TASF, 1.32 g, 4.8 mmol) at -15 °C, and then the
mixture was allowed to warm to room temperature. After 12 h, the
mixture was diluted with CH2Cl2, washed with water, and dried with
MgSO4. After concentration in vacuo, purification of the residue on a
silica gel column chromatography (toluene:ethyl acetate ) 3:1) gave
2 (R:â ) 1:24, 757 mg, 52%, 2 steps) as an amorphous mass; The
physical data were identical to the reported data.11a Conversion of 2
into 3 was then performed with reported procedure.11a
was neutralized by the addition of triethylamine. After concentration,
the mixture was dissolved in a solution of 0.3 M NaOH (14 mL) and
kept for 10 min at room temperature. The solution was made neutral
with IR120 (H
+) resin, and the mixture was filtered. After concentration,
purification of the residue on a gel permeation column chromatography
(Sephadex G-15, water) and subsequent lyophilization afforded 8 (R:â
) 1:15, 92 mg, 52%, 2 steps) as an amorphous mass; The physical
data were identical to the reported data.11d,28
Methyl (Methyl 5-Acetamido-4,7-di-O-benzyl-3,5-dideoxy-r,â-
D-glycero-D-galacto-2-nonulopyranosid) Onate (10). To a solution
of acetonide 911b (R:â ) 1:15, 3.2 g, 8.48 mmol) in dry N,N′-
dimethylformamide (43 mL) was added BaO (9.3 g, 60.65 mmol), Ba-
(OH)2‚8H2O (2.4 g, 7.61 mmol), and benzyl bromide (10 mL, 84.1
mmol), and this mixture was stirred at room temperature. After 12 h,
the mixture was diluted with CHCl3 and washed with aqueous 1%
formic acid and water. After drying with MgSO4, the organic solution
was evaporated. The residue was dissolved in ethanol (25 mL) and
benzene (50 mL), and then to this mixture was added an ethereal
solution of diazomethane (42.5 mmol). The mixture was stirred for 10
min at room temperature. After addition of acetic acid (12 mL), the
solution was concentrated in vacuo. The residue was dissolved in
aqueous 60% acetic acid (50 mL), and this mixture was stirred for 12
h at 60 °C. After concentration of this mixture, purification of the
residue on a silica gel column chromatography (ethyl acetate:MeOH
) 15:1) gave 10 (R:â ) 1:24, 2.7 g, 62%, 3 steps) as an amorphous
Methyl (Methyl 5-Acetamido-4,7,8-tri-O-benzyl-3,5,9-trideoxy-
r,â-D-glycero-D-galacto-2-nonulopyranosid) Onate (4). To a solution
of 1 (R:â ) 1:15, 637 mg, 1.05 mmol) in pyridine (5 mL) and CH2Cl2
(5 mL) was added phenyl chlorothionoformate (360 µL, 2.6 mmol) at
room temperature, and the mixture was stirred for 2 h at room
temperature. After addition of MeOH (1 mL), the mixture was
concentrated. The residue was diluted with ethyl acetate and washed
sequentially with saturated CuSO4, water (×2), aqueous NaHCO3, and
saturated NaCl. After drying with MgSO4, the organic phase was
concentrated. The residue was dissolved in toluene (55 mL), and to
this solution were added n-tributyltin hydride (2.8 mL, 10.4 mmol)
and 2,2′-azobisisobutyronitrile (AIBN, 52 mg, 0.32 mmol). The mixture
was stirred for 1 h at 80 °C and then concentrated in vacuo. Purification
of the residue on a silica gel column chromatography (toluene:ethyl
acetate ) 2:1) gave 4 (R:â ) 1:24, 240 mg, 39%, 2 steps) as an
1
mass; H NMR (D2O) δ 7.48-7.32 (m, 10H, Ar), 4.65, 4.59, 4.54,
4.44 (each d, each 1H, each J 11.9 Hz, 9.9 Hz, 9.9 Hz, 11.9 Hz, CH2-
Ph), 4.08 (dd, 1H, J4,5 10.6 Hz, J5,6 10.6 Hz, H-5), 3.91 (bd, 1H, J5,6
10.6 Hz, H-6), 3.91-3.82 (m, 2H, H-9a, H-8), 3.81 (m, 1H, J3a,4 11.2
Hz, J3e,4 4.6 Hz, J4,5 10.6 Hz, H-4), 3.79 (s, 3H, COOMe), 3.71 (m,
1H, H-9b), 3.70 (bd, 1H, J7,8 9.2 Hz, H-7), 3.21 (s, 3H, OMe), 2.51
(dd, 1H, J3e,4 4.6 Hz, Jgem 13.2 Hz, H-3e), 1.93 (s, 3H, NHAc), 1.68
(dd, 1H, J3a,4 11.2, Jgem 13.2 Hz, H-3a); 13C NMR (D2O) δ 174.32,
170.69, 137.90, 137.27, 129.36, 129.20, 129.13, 129.02, 128.98, 128.71,
99.66 (C-2), 76.50 (C-7), 75.33, 74.61 (C-4), 71.76, 71.02 (C-6), 70.30
(C-8), 63.13 (C-9), 53.94 (COOMe), 51.50 (OMe), 50.28 (C-5), 37.13
(C-3), 22.73 (COCH3); HRMS calcd for C27H35NO9 (M + Na+)
540.2210, found 540.2207.
1
amorphous mass; H NMR (CDCl3) δ 7.50-7.20 (m, 15H, Ar), 4.76
(s, 2H, CH2Ph), 4.71 (bd, 1H, J5,NH 8.6 Hz, NHAc), 4.64, 4.60, 4.51,
4.37 (each d, each 1H, each J 11.6 Hz, 12.2 Hz, 11.6 Hz, 12.2 Hz,
CH2Ph), 4.16 (ddd, 1H, J3a,4 10.6 Hz, J3e,4 4.6 Hz, J4,5 10.6 Hz, H-4),
4.14 (dd, 1H, J5,6 10.6 Hz, J6,7 1.3 Hz, H-6), 3.87 (dq, 1H, J7,8 4.6 Hz,
J8,9 5.9 Hz, H-8), 3.76 (s, 3H, COOMe), 3.66 (bdd, 1H, J6,7 1.3 Hz,
J7,8 4.6 Hz, H-7), 3.64 (ddd, 1H, J5,NH 8.6 Hz, J4,5 10.6 Hz, J5,6 10.6
Hz, H-5), 3.13 (s, 3H, OMe), 2.54 (dd, 1H, J3e,4 4.6 Hz, Jgem 13.2 Hz,
H-3e), 1.69 (s, 3H, NHAc), 1.69 (dd, 1H, J3a,4 10.6, Jgem 13.2 Hz, H-3a),
1.46 (d, 3H, J8,9 5.9 Hz, H-9); 13C NMR (CDCl3) δ 170.20, 168.37,
138.65, 138.54, 138.44, 128.99, 128.37, 128.35, 127.79, 127.77, 127.66,
127.48, 127.46, 99.00 (C-2), 77.42 (C-7), 76.86 (C-8), 73.19 (2C, C-4,
CH2Ph), 71.37 (C-6), 71.01, 70.51 (CH2Ph × 2), 52.82 (C-5), 52.47
(COOMe), 50.81 (OMe), 37.52 (C-3), 23.59 (COCH3), 15.94 (C-9);
HRMS calcd for C34H41NO8 (M + Na+) 614.2730, found 614.2742.
5-Acetamido-3,5,9-trideoxy-D-galacto-2-nonulopyranosonic Acid
(5). A solution of 4 (R:â ) 1:24, 187 mg, 0.32 mmol) in acetic acid
(3.5 mL) was stirred under hydrogen atmosphere in the presence of
10% Pd-C (187 mg) for 6 h. After filtration with Celite 545, the filtrate
was concentrated. The acetic acid in the residue was removed by
coevaporation with water in vacuo. The mixture was dissolved in a
solution of 0.3 M NaOH (7.5 mL) and MeOH (7.5 mL) and kept for
40 min at room temperature. The solution was made neutral with IR120
(H+) resin, and the mixture was filtered. After concentration, the residue
was dissolved in aqueous 0.025 M HCl (15 mL), and to this solution
was added Amberlyst 15 (wet) ion-exchange resin (1.1 g). Then the
mixture was stirred at 80 °C. After 4 h, the mixture was evaporated to
dryness and coevaporated with water (×3). The residue was dissolved
in water and passed through a gel permeation column chromatography
(Sephadex G-15, water). Fractions containing NeuAc derivative were
pooled and then the solution was lyophilized to give 5 (R:â ) 1:24,
78 mg, 64%, 3 steps); The physical data were identical to the reported
data.26
Methyl (Methyl 5-Acetamido-9-O-acetyl-4,7-di-O-benzyl-3,5-
dideoxy-r,â-D-glycero-D-galacto-2-nonulopyranosid) Onate (11). To
a solution of 10 (R:â ) 1:24, 398 mg, 0.77 mmol) in pyridine (4 mL)
and CH2Cl2 (2.4 mL) was added a solution of acetyl chloride (109 µL,
1.53 mmol) in CH2Cl2 (1.6 mL) at -45 °C, and the mixture was stirred
for 10 min at -45 °C. After addition of MeOH (1 mL), the solution
was concentrated in vacuo. Purification of the residue on a silica gel
column chromatography (ethyl acetate:MeOH ) 20:1) gave 11 (R:â
1
) 1:24, 354 mg, 82%) as an amorphous mass; H NMR (CDCl3) δ
7.44-7.25 (m, 10H, Ar), 4.96 (d, 1H, J5,NH 7.9 Hz, NHAc), 4.75 ∼
4.59 (m, 3H, CH2Ph), 4.48 (m, 1H, H-9a), 4.38 (d, 1H, CH2Ph), 4.26-
4.15 (m, 3H, H-6, H-9b, H-8), 4.06 (ddd, 1H, J3a,4 10.6 Hz, J3e,4 4.6
Hz, J4,5 9.9 Hz, H-4), 3.83 (m, 1H, H-5), 3.78 (s, 3H, COOMe), 3.61
(dd, 1H, J6,7 2.0 Hz, J7,8 5.3 Hz, H-7), 3.25 (s, 3H, OMe), 2.57 (dd,
1H, J3e,4 4.6 Hz, Jgem 13.2 Hz, H-3e), 2.07 (s, 3H, Ac), 1.74 (s, 3H,
Ac), 1.70 (dd, 1H, J3a,4 10.6 Hz, Jgem 13.2 Hz, H-3a); 13C NMR (CDCl3)
δ 171.37, 170.31, 168.12, 138.23, 137.92, 128.75, 128.43, 128.39,
127.92, 127.78, 127.71, 99.08 (C-2), 75.35 (C-7), 73.19 (C-4), 73.05,
70.89 (2C, C-6, CH2Ph), 70.06 (C-8), 66.47 (C-9), 52.60 (COOMe),
51.97 (C-5), 51.25 (OMe), 37.31 (C-3), 23.47, 20.86 (COCH3 × 2 );
Anal. Calcd for C29H37NO10: C, 62.24; H, 6.66, N, 2.50. Found: C,
61.94; H, 6.65; N, 2.79.
Methyl (Methyl 5-Acetamido-9-O-acetyl-4,7-di-O-benzyl-3,5,8-
trideoxy-r,â-D-galacto-2-nonulopyranosid) Onate (12). To a solution
of 11 (R:â ) 1:24, 945 mg, 1.69 mmol) in pyridine (8.5 mL) and CH2-
Cl2 (8.5 mL) was added 4-dimethylaminopyridine (104 mg, 0.85 mmol)
and phenyl chlorothionoformate (1.2 mL, 8.7 mmol), and the mixture
was stirred for 3 h at room temperature. After addition of MeOH (3
mL), the mixture was concentrated. The residue was dissolved in ethyl
acetate and washed with saturated NaCl. After drying with MgSO4,
the organic phase was concentrated in vacuo. The residue was dissolved
5-Acetamido-9-azido-3,5,9-trideoxy-D-glycero-D-galacto-2-nonulo-
pyranosonic Acid (8). Azido derivative 7 was prepared from known
compound 6 by 3 steps with reported procedure.11e To a solution of
this phenyl thioglycoside 711e (R:â ) 1:2, 233 mg, 0.53 mmol) in
acetone (4.8 mL) and H2O (0.5 mL) was added N-bromosuccinimide
(282 mg, 1.58 mmol) at room temperature.27 After 30 min, the mixture
(26) Zbiral, E.; Brandstetter, H. H.; Schreiner, E. P. Monatsh. Chem.
1988, 119, 127-141.
(27) Motawia, M. S.; Marcussen, J.; Møller, B. L. J. Carbohydr. Chem.
1995, 14, 1279-1294.
(28) Liu, J. L.-C.; Shen, G.-J.; Ichikawa, Y.; Rutan, J. F.; Zapata, G.;
Vann, W. F.; Wong, C.-H. J. Am. Chem. Soc. 1992, 114, 3901-3910.

Synthesis of Sialyloligosaccharide on Glycoprotein
J. Am. Chem. Soc., Vol. 122, No. 24, 2000 5691
in toluene (70 mL), and to this solution were added n-tributyltin hydride
(4.5 mL, 16.8 mmol) and AIBN (83 mg, 0.51 mmol). The mixture
was stirred for 30 min at 100 °C and then concentrated in vacuo.
Purification of this residue on a silica gel column chromatography
(toluene:ethyl acetate ) 1:1) gave 12 (R:â ) 1:24, 709 mg, 68%, 2
steps) as an amorphous mass; 1H NMR (CDCl3) δ 7.39-7.24 (m, 10H,
Ar), 4.88 (bd, 1H, NHAc), 4.66, 4.61, 4.52, 4.39 (each d, each 1H, J
11.9 Hz, CH2Ph), 4.33 ∼ 4.12 (m, 2H, H-9a, H-9b), 4.11 (ddd, 1H,
J3a,4 11.2 Hz, J3e,4 4.9 Hz, J4,5 9.9 Hz, H-4), 3.96 (dd, 1H, J6,7 2.3 Hz,
J5,6 10.2 Hz, H-6), 3.79 (s, 3H, COOMe), 3.73-3.62 (m, 2H, H-5,
H-7), 3.23 (s, 3H, OMe), 2.55 (dd, 1H, J3e,4 4.9 Hz, Jgem 13.2 Hz, H-3e),
2.13-2.03 (m, 2H, H-8a, 8b) 2.02 (s, 3H, NHAc), 1.71 (dd, 1H, J3a,4
11.2 Hz, Jgem 13.2 Hz, H-3a); 13C NMR (CDCl3) δ 171.01, 170.10,
168.32, 138.38, 138.28, 128.61, 128.48, 128.39, 127.91, 127.78, 127.69,
99.10 (C-2), 73.26 (2C, C-7, C-4), 71.93 (C-6), 71.84, 71.11, 61.46
(C-9), 52.78 (C-5), 52.54 (COOMe), 50.93 (OMe), 37.72 (C-3), 29.31
(C-8), 23.52, 20.95 (COCH3 × 2); Anal. Calcd for C29H37NO9: C,
64.07; H, 6.86, N, 2.58. Found: C, 63.91; H, 6.58; N, 2.86.
and AIBN (15 mg, 0.09 mmol). The mixture was stirred for 30 min at
80 °C and then concentrated. Purification of the residue on a silica gel
column chromatography (toluene:ethyl acetate ) 1:1) gave 15 (R:â )
1:19, 111 mg, 80%, 2 steps) as an amorphous mass; 1H NMR (CDCl3)
δ 8.00-7.35 (m, 5H, Ar), 5.64 (d, J5,NH 9.4 Hz, 1H, NHAc), 5.46 (ddd,
1H, J3a,4 11.1 Hz, J3e,4 5.1 Hz, J4,5 11.1 Hz, H-4), 4.35 (m, 1H, H-8),
4.12 (ddd, 1H, J5,NH 9.4 Hz, J4,5 11.1 Hz, J5,6 10.3 Hz, H-5), 4.10 (dd,
1H, J8,9a 5.9 Hz, Jgem 7.8 Hz, H-9a), 3.85 ∼ 3.75 (m, 4H, COOMe,
H-6), 3.56 (dd, 1H, J8,9b 7.8 Hz, Jgem 7.8 Hz, H-9b), 3.31 (s, 3H, OMe),
2.55 (dd, 1H, J3e,4 5.1 Hz, Jgem 12.7 Hz, H-3e), 2.01 (dd, 1H, J3a,4 11.1
Hz, Jgem 12.7 Hz, H-3a), 1.95-1.82 (m, 5H, NHAc, H-7a, H-7b); 13C
NMR (CDCl3) δ 170.41, 168.09, 166.81, 133.38, 129.72, 129.38,
128.46, 108.53, 98.38 (C-2), 72.53 (C-8), 70.38 (C-6), 69.75 (C-9),
69.51 (C-4), 53.98 (C-5), 52.64 (COOMe), 50.92 (OMe), 37.75 (C-3),
35.71 (C-7), 27.07, 25.76 (Me × 2), 23.23 (COCH3); HRMS calcd for
C23H31NO9 (M + H+) 466.2077, found 466.2065.
5-Acetamido-3,5,7-trideoxy-D-galacto-2-nonulopyranosonic Acid
(16). Acetonide 15 was (R:â ) 1:19, 303 mg, 0.65 mmol) dissolved in
aqueous 60% acetic acid (6.2 mL), and this mixture was stirred for 30
min at 60 °C. After concentration, the mixture was dissolved in a
solution of 0.3 M NaOH (15 mL) and MeOH (15 mL), and kept for 10
min at room temperature. The solution was made neutral with IR120
(H+) resin, and the mixture was filtered. After concentration of the
filtrate, the residue was dissolved in aqueous 0.025 M HCl (38 mL),
and then to this solution was added Amberlyst 15 (wet) ion-exchange
resin (3.0 g). The mixture was stirred for 4 h at 80 °C. After filtration
of the reaction mixture, the filtrate was evaporated to dryness and
coevaporated with water (×3). The residue was dissolved in water and
passed through a gel permeation column chromatography (Sephadex
G-15, water). Fractions containing NeuAc derivative were pooled, and
the resulting solution was lyophilized to give 16 (R:â ) 1:11, 132 mg,
69%, 3 steps); the physical data were identical to the reported data.28
5-Acetamido-3,5,8-trideoxy-D-galacto-2-nonulopyranosonic Acid
(13). A solution of 12 (R:â ) 1:24, 607 mg, 1.12 mmol) in acetic acid
(16 mL) was stirred under hydrogen atmosphere in the presence of
10% Pd-C (598 mg) for 2 h. After filtration with Celite 545, the filtrate
was concentrated. The acetic acid in the residue was removed by
coevaporation with water in vacuo. The mixture was dissolved in 0.3
M NaOH (7.4 mL) and MeOH (3.8 mL), and kept for 1 h at room
temperature. The solution was made neutral with IR120 (H+) resin,
and the mixture was filtered. After concentration, the residue was
dissolved in aqueous 0.025 M HCl (38 mL), and to this solution was
added Amberlyst 15 (wet) ion-exchange resin (1.9 g). Then the mixture
was stirred at 85 °C. After 4 h, the mixture was evaporated to dryness
and coevaporated with water (×3). The residue was dissolved in water
and passed through a column of anion-exchange resin (AG1-X8, HCO2-
form). After eluting of impurities by water, derivative 13 was eluted
by aqueous 1 M formic acid, and the eluant was evaporated. The formic
acid in the residue was removed by coevaporation with water in vacuo.
The residue was lyophilized to give 13 (R:â ) 1:24, 193 mg, 59%, 3
steps); The physical data were identical to the reported data.26,29
Methyl (Methyl 5-Acetamido-4-O-benzoyl-3,5-dideoxy-8,9-O-iso-
propylidene-r,â-D-glycero-D-galacto-2-nonulopyranosid) Onate (14).
To a solution of acetonide 911b (R:â ) 1:15, 1.0 g, 2.8 mmol) in pyridine
(6 mL) and CH2Cl2 (24 mL) was added benzoyl chloride (974 µL, 8.4
mmol) dropwise at - 5 °C. The mixture was stirred for 1 h at 0 °C,
and then to this mixture was added MeOH (200 µL). After concentra-
tion, purification of the residue on a silica gel column chromatography
(toluene:ethyl acetate ) 2:1) gave 14 (R:â ) 1:32, 908 mg, 68%) as
Monitoring of [3-13C]-Labeling for NeuAc Analogues by Ex-
change Reaction of NeuAc Aldolase. NeuAc analogue (15 µmol) was
dissolved in a solution of phosphate buffer (150 mM, pH 7.0, total
500 µL) containing NaN3 (0.1 mg), NeuAc aldolase (10 U), [3-13C]-
sodium pyruvate (45 µmol), and D2O (20 µL). The reaction was carried
out in NMR sample tube. The quantity of [3-13C]-NeuAc analogue and
[3-13C]-pyruvic acid were determined by 1H NMR measurement
(monitoring times: 0, 4, 8, 16, 24, 40, 48, and 96 h, scans: 256, temp.
30 °C). After 96 h, the reaction mixture was passed through a column
-
of anion-exchange resin (AG1-X8, HCO2 form). The impurities was
removed by eluting with water, and then NeuAc analogue was eluted
with aqueous 1 M formic acid. A solution containing [3-13C]-NeuAc
analogue was concentrated. The formic acid in the residue was removed
by coevaporation with water in vacuo. 13C-labeling yield was determined
1
an amorphous mass; H NMR (CD3OD) δ 8.10-7.40 (m, 5H, Ar),
1
5.49 (ddd, 1H, J3a,4 11.2 Hz, J3e,4 5.3 Hz, J4,5 10.6 Hz, H-4), 4.39 (dd,
1H, J4,5 10.6 Hz, J5,6 10.6 Hz, H-5), 4.32 (ddd, 1H, J7,8 8.6 Hz, J8,9a 5.9
Hz, J8,9b 5.9 Hz, H-8), 4.16 (dd, 1H, J8,9a 5.9 Hz, Jgem 8.3 Hz, H-9a),
4.00 (dd, 1H, J8,9b 5.9 Hz, Jgem 8.3 Hz, H-9b), 3.98 (dd, 1H, J5,6 10.6
Hz, J6,7 1.3 Hz, H-6), 3.84 (s, 3H, COOMe), 3.57 (dd, 1H, J6,7 1.3 Hz,
J7,8 8.6 Hz, H-7), 3.35 (s, 3H, OMe), 2.63 (dd, 1H, J3e,4 5.3 Hz, Jgem
12.9 Hz, H-3e), 1.91 (dd, 1H, J3a,4 11.2 Hz, Jgem 12.9 Hz, H-3a), 1.89
(s, 3H, NHAc), 1.40, 1.33 (each s, each 3H, Me × 2); 13C NMR (CD3-
OD) δ 174.65, 171.00, 168.15, 135.23, 131.81, 131.45, 130.39, 111.17,
100.96 (C-2), 76.72 (C-8), 73.34 (C-6), 72.46 (C-4), 72.10 (C-7), 69.42
(C-9), 54.11 (COOMe), 52.46 (OMe), 51.09 (C-5), 39.25 (C-3), 28.09,
26.44 (Me × 2), 23.42 (COCH3); Anal. Calcd for C23H31NO10: C,
57.37; H, 6.49, N, 2.91. Found: C, 57.08; H, 6.60; N, 3.05.
by H NMR and was summarized in Table 1. NeuAc (4.6 mg, 14.9
µmol), 9-deoxy-9-fluoro-NeuAc 3 (4.7 mg, 15.1 µmol) 9-deoxy-NeuAc
5 (4.4 mg, 15.0 µmol), 9-azido-9-deoxy-NeuAc 8 (5.4 mg, 15.1 µmol),
8-deoxy-NeuAc 13 (4.4 mg, 15.0 µmol), and 7-deoxy-NeuAc 16 (4.4
mg, 15.0 µmol) were examined by this exchange reaction with the
conditions described above.
Kinetic Measurements toward NeuAc Aldolase. The assay used
was based on the method described.12 NeuAc analogues or NeuAc were
incubated in a solution of phosphate buffer (50 mM, pH 7.0, total
volume 80 µL) containing 0.2 mM â-NADH, 25 U of LDH, and 4 mU
of NeuAc aldolase. In the case of 8-deoxy-NeuAc 13 and 7-deoxy-
NeuAc 16, 20 mU of NeuAc aldolase was used. The mixture was
incubated until less 1% consumption of starting material at 37 °C. The
amount of pyruvic acid released in the assay mixture was estimated
by measurement of decreasing of absorbance at 340 nm. All assays
were performed in duplicate with five concentrations of NeuAc substrate
(NeuAc and 9-modified NeuAc: 1.25, 2.5, 5.0, 7.5, and 10 mM;
8-deoxy-NeuAc: 2.5, 5.0, 7.5, 10, and 15 mM, 7-deoxy-NeuAc: 5.0,
7.5, 10, 12.5, and 15 mM). The kinetic parameters Km and Vmax values
were obtained from the Lineweaver-Burk plots.
Methyl (Methyl 5-Acetamido-4-O-benzoyl-3,5,7-trideoxy-8,9-O-
isopropylidene-r,â-D-glycero-D-galacto-2-nonulopyranosid) Onate
(15). To a solution of 14 (R:â ) 1:32, 144 mg, 0.30 mmol) in CH2Cl2
(6 mL) was added 4-dimethylaminopyridine (18 mg, 0.15 mmol) and
1,1′-thiocarbonyldiimidazole (160 mg, 0.90 mmol) at room temperature,
and the mixture was stirred for 12 h under the reflux condition. After
addition of MeOH (100 µL), the mixture was diluted with ethyl acetate
and washed with water (×2). After drying with MgSO4, the mixture
was concentrated. The residue was dissolved in toluene (23 mL), and
to this solution were added n-tributyltin hydride (0.4 mL, 1.49 mmol)
Degradation Reaction by the Use of NeuAc Aldolase and Lactate
Dehydrogenase. NeuAc analogue (30 mM) was dissolved in a solution
of phosphate buffer (0.5 M, pH 7.5) containing BSA (0.1 mg), NaN3
(0.1 mg), NeuAc aldolase (3 U), â-NADH (45 mmol), and LDH (9.3
U), and the mixture was incubated at 37 °C. The quantity of NeuAc
(29) David, S.; Malleron, A.; Cavaye´, B. New J. Chem. 1992, 16, 751-
755.

5692 J. Am. Chem. Soc., Vol. 122, No. 24, 2000
Miyazaki et al.
analogue in the reaction mixture was determined by thiobarbituric acid
assay13 (TBA, monitoring times: 0, 5, 15, and 45 min and 1.75, 3.75,
5.75, and 9.75 h). 7-Deoxy-NeuAc 16 could not be assayed by TBA
because both 4-deoxy-ManNAc formed and 7-deoxy-NeuAc were
positive in this assay. Therefore, the quantity of 7-deoxy-NeuAc was
determined by HPLC equipping a column chromatography of amino-
propyl (0.4 cm × 25 cm, Asahipak, NH2-P50), which was run with a
mixture of CH3CN/15 mM KH2PO4 ) 7/4. 7-Deoxy-NeuAc was
detected at 200 nm, and its retention time was found to be 14.3 min.
NeuAc (1.9 mg, 6.1 µmol), 9-deoxy-9-fluoro-NeuAc 3 (1.5 mg, 4.8
µmol), 9-deoxy-NeuAc 5 (1.5 mg, 5.1 µmol), 9-azido-9-deoxy-NeuAc
8 (1.2 mg, 3.3 µmol), 8-deoxy-NeuAc 13 (1.0 mg, 3.4 µmol), and
7-deoxy-NeuAc 16 (2.0 mg, 6.8 µmol) were examined by this
degradation reaction with the conditions described above. Enzyme assay
suggested that all degradation yields of NeuAc and NeuAc analogues
3, 5, 8, 13, and 16 were quantitative.
(d, JC8,F 17.7 Hz, C-8), 67.58 (d, JC7,F 7.2 Hz, C-7), 67.15 (d, JC3,C4
35.9 Hz, C-4), 52.46 (C-5), 39.26 (C-3), 22.51 (COCH3); HRMS calcd
for C1013CH18FNO8 (M + H+) 313.1128, found 313.1150.
[3-13C]-5-Acetamido-3,5,9-trideoxy-D-galacto-2-nonulopyranoson-
ic Acid ([3-13C]-9-deoxy-NeuAc). R:â ) 1:15; 1H NMR (D2O) δ 4.08
(m, 1H, J3a,4 12.8 Hz, J3e,4 4.7 Hz, J4,5 10.1 Hz, H-4), 4.08 (bd, 1H, J5,6
10.3 Hz, H-6), 3.99 (dd, 1H, J4,5 10.1 Hz, J5,6 10.3 Hz, H-5), 3.92 (dq,
1H, J7,8 8.3 Hz, J8,9 6.4 Hz, H-8), 3.39 (bd, 1H, J7,8 8.3 Hz, H-7), 2.28
(ddd, 1H, J3e,4 4.7 Hz, Jgem 12.8 Hz, J3e,C3 132.3 Hz, H-3e), 2.13 (s,
3H, NHAc), 1.91 (ddd, 1H, J3a,4 12.8 Hz, Jgem 12.8 Hz, J3a,C3 129.1
Hz, H-3a), 1.32 (d, 3H, J8,9 6.4 Hz, H-9); 13C NMR (D2O) δ 177.07,
175.11, 96.79 (d, JC2,C3 41.2 Hz, C-2), 73.17 (C-7), 70.61 (C-6), 67.49
(d, JC3,C4 35.3 Hz, C-4), 67.06 (C-8), 52.89 (C-5), 39.78 (C-3), 22.55
(COCH3), 19.73 (C-9); HRMS calcd for C1013CH19NO8 (M + Na+)
317.1042, found 317.1032.
[3-13C]-5-Acetamido-9-azido-3,5,9-trideoxy-D-glycero-D-galacto-
2-nonulo-pyranosonic Acid ([3-13C]-9-azido-9-deoxy-NeuAc). R:â )
1:15; 1H NMR (D2O) δ 4.13 (m, 1H, J3a,4 11.9 Hz, J3e,4 4.6 Hz, J4,5 9.9
Hz, H-4), 4.12 (bd, 1H, J5,6 10.6 Hz, H-6), 4.00 (dd, 1H, J4,5 9.9 Hz,
J5,6 10.6 Hz, H-5), 3.99 (ddd, 1H, J7,8 9.2 Hz, J8,9a 5.9 Hz, J8,9b 2.6 Hz,
H-8), 3.70 (dd, 1H, J8,9a 2.6 Hz, Jgem 13.2 Hz, H-9a), 3.64 (bd, 1H, J7,8
9.2 Hz, H-7), 3.55 (dd, 1H, J8,9b 5.9 Hz, Jgem 13.2 Hz, H-9b), 2.36
(ddd, 1H, J3e,4 4.6 Hz, Jgem 13.2 Hz, J3e,C3 133.3 Hz, H-3e), 2.14 (s,
3H, NHAc), 1.95 (ddd, 1H, J3a,4 11.9 Hz, Jgem 13.2 Hz, J3a,C3 129.3
Hz, H-3a); 13C NMR (D2O) δ 175.27, 174.23, 95.92 (d, JC2,C3 41.6 Hz,
C-2), 70.65 (C-6), 69.45 (C-8), 69.19 (C-7), 67.21 (d, JC3,C4 35.8 Hz,
C-4), 54.30 (C-9), 52.56 (C-5), 39.37 (C-3), 22.52 (COCH3); HRMS
calcd for C1013CH18N4O8 (M + H+) 336.1236, found 336.1242.
[3-13C]-5-Acetamido-3,5,8-trideoxy-D-galacto-2-nonulopyranoson-
ic Acid ([3-13C]-8-deoxy-NeuAc). R:â ) 1:24; 1H NMR (D2O) δ 4.08
(m, 1H, J3a,4 11.9 Hz, J3e,4 4.6 Hz, J4,5 9.9 Hz, H-4), 3.96 (dd, 1H, J4,5
9.9 Hz, J5,6 10.6 Hz, H-5), 3.90 (ddd, 1H, J6,7 1.3 Hz, J7,8a 8.6 Hz, J7,8b
4.6 Hz, H-7), 3.77 (dd, 2H, J8a,9 6.6 Hz, J8b,9 6.6 Hz, H-9), 3.70 (dd,
1H, J5,6 10.6 Hz, J6,7 1.3 Hz, H-6), 2.32 (ddd, 1H, J3e,4 4.6 Hz, Jgem
12.5 Hz, J3e,C3 132.6 Hz, H-3e), 2.14 (s, 3H, NHAc), 1.99 (tdd, 1H,
J7,8a 8.6 Hz, Jgem 14.5 Hz, J8a,9 6.6 Hz, H-8a), 1.91 (ddd, 1H, J3a,4 11.9
Hz, Jgem 12.5 Hz, J3a,C3 129.3 Hz, H-3a), 1.78 (tdd, 1H, J7,8b 4.6 Hz,
Jgem 14.5 Hz, J8b,9 6.6 Hz, H-8b); 13C NMR (D2O) δ 175.39, 173.81,
95.65 (d, JC2,C3 41.2 Hz, C-2), 74.18 (C-6), 67.08 (d, JC3,C4 35.7 Hz,
C-4), 65.88 (C-7) 58.97 (C-9), 53.07 (C-5), 39.39 (C-3), 35.33 (C-8),
22.48 (COCH3); HRMS calcd for C1013CH19NO8 (M + H+) 295.1223,
found 295.1221.
[3-13C]-5-Acetamido-3,5,7-trideoxy-D-galacto-2-nonulopyranoson-
ic Acid ([3-13C]-7-deoxy-NeuAc). R:â ) 1:11; 1H NMR (D2O) δ 4.07
(ddd, 1H, J5,6 10.0 Hz, J6,7a 2.7 Hz, J6,7b 10.0 Hz, H-6), 4.06 (m, 1H,
H-4), 3.93 (dddd, 1H, J7a,8 6.4 Hz, J7b,8 6.4 Hz, J8,9a 3.8 Hz, J8,9b 6.8
Hz, H-8), 3.68 (dd, 1H, J4,5 10.0 Hz, J5,6 10.0 Hz, H-5), 3.65 (dd, 1H,
J8,9a 3.8 Hz, Jgem 11.7 Hz, H-9a), 3.52 (dd, 1H, J8,9b 6.8 Hz, Jgem 11.7
Hz, H-9b), 2.39 (ddd, 1H, J3e,4 4.9 Hz, Jgem 13.1 Hz, J3e,C3 133.4 Hz,
H-3e), 2.16 (s, 3H, NHAc), 1.97 (ddd, 1H, J3a,4 12.5 Hz, Jgem 13.1 Hz,
J3a,C3 129.7 Hz, H-3a), 1.78-1.63 (m, 2H, H-7a, H-7b); 13C NMR (D2O)
δ 175.17, 173.65, 95.39 (d, JC2,C3 41.7 Hz, C-2), 69.31 (C-6), 68.29
(C-8), 66.99 (d, JC3,C4 35.7 Hz, C-4), 66.20 (C-9), 56.65 (C-5), 39.39
(C-3), 34.80 (C-7), 22.58 (COCH3); HRMS calcd for C1013CH19NO8
(M + H+) 295.1223, found 295.1225.
Typical Procedure of a One-Pot Enzymatic 13C-Labeling Method.
NeuAc analogue (30 mM) was dissolved in a solution of phosphate
buffer (50 mM, pH 7.5) containing BSA (5 mg), NaN3 (5 mg), NeuAc
aldolase (30 U), â-NAD+ (0.3 mM), LDH (150 U), ADH (150 U),
and EtOH (120 µL, 2.1 mmol), and then the pH was adjusted to 7.5.
The mixture was incubated at room temperature for 20 h. After
completion of the degradation reaction (ascertained by the TBA), to
this mixture was added NPP (2.8 U) and MgCl2 (6.0 mg). The mixture
was incubated at room temperature for 6 h. This NPP reaction was
monitored by TLC (solvent system: ethyl acetate/MeOH/H2O ) 3/2/
1.5; Rf â-NAD+ ) 0.20, â-NADH ) 0.45). After completion of the
degradation of â-NADH and â-NAD+, to this solution was added
[3-13C]-sodium pyruvate (3 equiv toward NeuAc analogue), and the
mixture was incubated at room temperature for 24 h. After lyophiliza-
tion, the residue was passed through a column of anion-exchange resin
(AG1-X8, HCO2 form). After elution of impurities by water, 13C-
-
labeled NeuAc analogue was eluted with aqueous 1 M formic acid,
and the eluant was evaporated. The formic acid in the solution was
removed by coevaporation with water. Purification of the residue on a
gel permeation column chromatography (Sephadex G-15, H2O) afforded
[3-13C]-NeuAc analogue. The quantity of NeuAc analogue during the
reaction was monitored by TBA (monitoring time: 0, 15, 45 min and
1.75, 3.75, 7.75, 15.75, 19.75, 23.75, 26.05, 26.80, 27.80, 29.80, 33.80,
41.80, and 49.80 h). In the case of 7-deoxy-NeuAc 16, the compound
was monitored by HPLC. These results of monitoring are shown in
Figure 6. NeuAc (30.9 mg, 0.1 mmol), 9-deoxy-9-fluoro-NeuAc 3 (31
mg, 0.10 mmol), 9-deoxy-NeuAc 5 (30.5 mg, 0.1 mmol), 9-azido-9-
deoxy-NeuAc 8 (30.1 mg, 0.09 mmol), 8-deoxy-NeuAc 13 (31.3 mg,
0.11 mmol), and 7-deoxy-NeuAc 16 (33.2 mg, 0.11 mmol) were
examined by this one-pot 13C-labeling method with condition described
above. The isolated yields and 13C-labeling yields are summarized in
Table 2.
[3-13C]-5-Acetamido-3,5-dideoxy-D-glycero-D-galacto-2-nonulo-
1
pyranosonic Acid ([3-13C]-NeuAc). R:â ) 1:15; H NMR (D2O) δ
4.12 (m, 1H, J3a,4 11.2 Hz, J3e,4 4.6 Hz, J4,5 9.9 Hz, H-4), 4.10 (bd, 1H,
J5,6 9.9 Hz, H-6), 4.00 (dd, 1H, J4,5 9.9 Hz, J5,6 9.9 Hz, H-5), 3.92 (dd,
1H, J8,9a 2.6 Hz, Jgem 11.2 Hz, H-9a), 3.84 (ddd, 1H, J7,8 8.6 Hz, J8,9a
5.9 Hz, J8,9b 2.6 Hz, H-8), 3.70 (dd, 1H, J8,9b 5.9 Hz, Jgem 11.2 Hz,
H-9b), 3.63 (bd, 1H, J7,8 8.6 Hz, H-7), 2.35 (ddd, 1H, J3e,4 4.6 Hz, Jgem
12.5 Hz, J3e,C3 133.3 Hz, H-3e), 2.13 (s, 3H, NHAc), 1.94 (ddd, 1H,
J3a,4 11.2 Hz, Jgem 12.5, J3a,C3 130.0 Hz, H-3a); 13C NMR (D2O) δ
175.71, 175.27, 96.56 (d, JC2,C3 41.2 Hz, C-2), 70.97 (C-8), 70.86 (C-
6), 68.67 (C-7), 67.35 (d, JC3,C4 35.5 Hz, C-4), 63.51 (C-9), 52.58 (C-
5), 39.57 (C-3), 22.53 (COCH3); HRMS calcd for C1013CH19NO9 (M
+ H+) 311.1172, found 311.1185.
[3-13C]-Methyl (5-Acetamido-4,7,8,9-tetra-O-acetyl-3,5-dideoxy-
r,â-D-glycero-D-galacto-2-nonulopyranosonic Acid (18). A mixture
of [3-13C]-NeuAc 17 (R:â ) 1:15, 184 mg, 0.59 mmol) and dry Dowex
50W-X8 (H+, 500 mg) in anhydrous MeOH (40 mL) was stirred for 3
h at room temperature to give the methyl ester (165 mg). This methyl
ester was used in the next step without purification. To a solution of
the methyl ester in acetic anhydride (1.5 mL) was added aqueous 60%
HClO4:Ac2O ) 1:9 (33 µL) at 0 °C.16 The mixture was stirred for 6 h
at 10 °C. The acetate derivative was diluted with ethyl acetate, and the
organic solution was washed with water and then saturated aqueous
NaHCO3. After drying with MgSO4, the solution was evaporated.
Purification of the residue on silica gel column chromatography (ethyl
acetate) gave 18 (R:â ) 1:13; 109 mg, 37%, 2 steps) as an amorphous
[3-13C]-5-Acetamido-3,5,9-trideoxy-9-fluoro-D-glycero-D-galacto-
2-nonulo-pyranosonic Acid ([3-13C]-9-deoxy-9-fluoro-NeuAc). R:â
) 1:19; 1H NMR (D2O) δ 4.74 (dd, 2H, J8,9 3.3 Hz, J9,F 46.1 Hz, H-9),
4.12 (m, 1H, J3a,4 11.2 Hz, J3e,4 4.6 Hz, J4,5 9.7 Hz, H-4), 4.08 (bd, 1H,
J5,6 9.9 Hz, H-6), 4.01 (dd, 1H, J4,5 9.7 Hz, J5,6 9.9 Hz, H-5), 4.00 (tdd,
1H, J7,8 9.2 Hz, J8,9 3.3 Hz, J8,F 27.7 Hz, H-8), 3.71 (bd, 1H, J7,8 9.2
Hz, H-7), 2.32 (ddd, 1H, J3e,4 4.6 Hz, Jgem 12.5 Hz, J3e,C3 132.6 Hz,
H-3e), 2.13 (s, 3H, NHAc), 1.93 (ddd, 1H, J3a,4 11.2 Hz, Jgem 12.5 Hz,
J3a,C3 131.3 Hz, H-3a); 13C NMR (D2O) δ 175.23, 173.66, 95.74 (d,
JC2,C3 40.6 Hz, C-2), 85.56 (d, JC9,F 164.9 Hz, C-9), 70.69 (C-6), 69.14
1
mass; H NMR (CDCl3) δ 5.69 (bd, 1H, J5,NH 9.5, NHAc), 5.35 (dd,
1H, J6,7 1.8 Hz, J7,8 5.6 Hz, H-7), 5.24 (ddd, 1H, J7,8 5.6 Hz, J8,9a 2.5
Hz, J8,9b 7.5 Hz, H-8) 5.27-5.17 (m, 1H, J3a,4 11.4 Hz, J3e,4 5.5 Hz,

Synthesis of Sialyloligosaccharide on Glycoprotein
J. Am. Chem. Soc., Vol. 122, No. 24, 2000 5693
J4,5 10.0 Hz, H-4), 4.50 (dd, 1H, J8,9a 2.5 Hz, Jgem 12.3 Hz, H-9a), 4.22
(bdd, 1H, J5,6 10.2 Hz, J6,7 1.8 Hz, H-6), 4.16 (ddd, 1H, J4,5 10.0 Hz,
J5,6 10.2 Hz, J5,NH 9.5 Hz, H-5), 4.02 (dd, 1H, J8,9b 7.5 Hz, Jgem 12.3
Hz, H-9b), 3.86 (s, 3H, COOMe), 2.26 (ddd, 1H, J3a,4 11.4 Hz, Jgem
12.7 Hz, J3a,C3 127.7 Hz, H-3a) 2.19 (ddd, 1H, J3e,4 5.5 Hz, Jgem 12.7
Hz, J3e,C3 134.1 Hz, H-3e), 2.14, 2.09, 2.03, 2.02, 1.90 (each s, each
3H, NHAc, Ac × 4); 13C NMR (CDCl3) δ 170.98, 170.88, 170.82,
170.31, 170.20, 169.07, 94.85 (d, JC2,C3 43.0 Hz, C-2), 71.22 (C-8),
71.02 (C-6), 69.31 (d, JC3,C4 36.4 Hz, C-4), 68.01 (C-7), 62.65 (C-9),
53.45 (COOMe), 49.52 (C-5), 36.11 (C-3), 23.12, 20.99, 20.88, 20.76,
20.76 (NHAc, Ac × 4); HRMS calcd for C1913CH29NO13 (M + H+)
493.1751, found 493.1760.
µL) containing UDP-[U-13C]-Glc (15.0 mg, 25.2 µmol), methyl
2-acetamido-2-deoxy-â-D-glucoside 24 (5.0 mg, 21.3 µmol), UDP-
glucose-4-epimerase (540 mU), boVine â-1,4-galactosyltransferase (270
mU), MnCl2 (50 mM), CIAP (50 U). The mixture was incubated at 37
°C for 36 h. Purification of the mixture on a silica gel column
chromatography (ethyl acetate:MeOH ) 3:2) gave 25 (4.7 mg, 57%);
The NMR data (Table 3) are identical to unlabeled Gal-â-(1f4)-
GlcNAc-â-OMe.30
[3-13C]-NeuAc-r-(2f3)-[U-13C]-Gal-â-(1f4)-GlcNAc-â-OMe (26).
The transfer reaction was performed in a solution of sodium cacodylate
buffer (50 mM, pH 6.0, 360 µL) containing CMP-[3′′-13C]-NeuAc 20
(3.6 mg, 5.8 µmol), [U-13C]-Gal-â-(1f4)-GlcNAc-â-OMe 25 (2.1 mg),
CIAP (25U), rat recombinant R-2,3-(N)-sialyltransferase (50 mU), and
the mixture was incubated at 37 °C for 24 h. The mixture was passed
through a gel permeation column chromatography (Sephadex G-15,
water). Then the residue was purified by HPLC equipping a column
chromatography of aminopropyl (0.4 cm × 25 cm, Asahipak, NH2-
P50. retention time ) 11.0 min), which was run with a mixture of
CH3CN/15 mM KH2PO4 ) 7/3. Further HPLC purification using a gel
permeation column chromatography (0.8 cm × 50 cm, Asahipak, GS-
320, H2O, retention time ) 29.6 min) gave 26 (1.1 mg, 30%); The
NMR data (Table 3) are identical to unlabeled NeuAc-R-(2f3)-Gal-
â-(1f4)-GlcNAc-â-OMe.30
NMR Methods for Conformational Analysis. For NMR studies,
the glycoprotein (40-50 mg) which was dialyzed with Molcut-L filter
was dissolved in 300 µL of D2O. All NMR experiments were run on
a Bruker AVANCE 400 instrument. The sample was set to 303 K.
Two-dimensional 1H-13C HMQC, HMQC-TOCSY, and HMQC-
NOESY spectra were measured by use of pulse programs in the Bruker
standard library (invbtp, invbmltp, and invbnotp, respectively). During
acquisition, GARP decoupling was performed toward 13C (F1 dimen-
sion). Pulse sequences of 1D HMQC (selective excitation of H-3axNeuAc
or H-3eqNeuAc), 1D HMQC-NOESY(selective excitation of H-3axNeuAc
or H-3eqNeuAc), 1D HSQC-TOCSY-NOESY, and 1D HSQC-TOCSY-
NOESY-TOCSY were made with combined Bruker standard programs
and a pulse sequence of TOCSY-NOESY-TOCSY10 modified as shown
in Figure 13. The narrow bars, wide bars, and wide bars having diagonal
lines represent 90°, 180°, and trim pulse, respectively. The trim pulse
at the end of MLEV17 was set for 2.5 ms, and the other trim pulse
was set for 1 ms. (A) 1D HMQC experiment with Gaussian (90°, 7
ms) and half-Gaussian (90°, 1 ms) selective pulses for H-3axNeuAc (or
H-3eqNeuAc) and C-3NeuAc, respectively. The following delay time and
phase cycling were applied: d1 (2.5 s), d2 (3.84 ms), d3 (484 ms), d4
(336 µs), d5 (3 µs); φ1 ) x, φ2 ) -x, φ3 ) x, φ4 ) x, φ5 ) 8x,
8(-x), φ6 ) x, y, -x, -y, φ7 ) 4x, 4(-x), receiver phase ) x, -y,
-x, y, -x, y, x, -y. (B) 1D HMQC-NOESY experiment with Gaussian
(90°, 7 ms) and half-Gaussian (90°, 1 ms) selective pulses for H-3axNeuAc
(or H-3eqNeuAc) and C-3NeuAc, respectively. The following delay time
and phase cycling were applied: d1 (2.5 s), d2 (3.84 ms), d3 (484 ms),
d4 (336 µs), d5 (3 µs), d6 (mixing time); φ1 ) 16x, 16(-x), φ2 ) -x,
φ3 ) x, φ4 ) x, φ5 ) 32x, 32(-x), φ6 ) x, -x, φ7 ) 2x, 2(-x), φ8
) x, φ9 ) 4x, 4y, 4(-x), 4(-y), φ10 ) x, receiver phase ) x, 2(-x),
x, y, 2(-y), y, -x, 2x, -x, -y, 2y, -y, -x, 2(x), -x, -y, 2y, -y, x,
2(-x), x, y, 2(-y), y. (C) 1D HSQC-TOCSY experiment. Dulations
for E-BURP (90°) for C3NeuAc selective pulses31 was 4 ms. Spin locking
(MLEV17) time was 50 ms. The following delay time and phase cycling
were applied: d1 (1 s), d2 (1.6 ms), d3 (3 µs), d4 (1.1 ms), d5 (3 µs),
d6 (100 µs), d7 (26 µs), d8 (3 µs), d9 (1.05 ms); φ1 ) x, φ2 ) x, φ3
) y, φ4 ) 2x, 2(-x), φ5 ) x, φ6 ) x, -x, φ7 ) y, -y, φ8 ) x, φ9 )
-y, φ10 ) x, φ11 ) 4x, 4(x), φ12 ) x, φ13 ) x, φ14 ) y, φ15 ) x,
φ26 ) y, receiver phase ) x, -x, x, -x, -x, x, -x, x. The following
gradient ratios were calculated and then used: gz1:gz2:gz3 ) 40:15:
10.05. (D) 1D HSQC-TOCSY-NOESY experiment. Dulations for
E-BURP (90°) for C-3NeuAc selective pulses31 was 4 ms. The following
delay time and phase cycling were applied: d1 (1 s), d2 (1.6 ms), d3
(3 µs), d4 (1.1 ms), d5 (3 µs), d6 (100 µs), d7 (26 µs), d8 (3 µs), d9
CMP-[3′′-13C]-NeuAc (20). The acetate derivative 18 (R:â ) 1:13,
86.5 mg, 0.176 mmol) and cytidine 5′-O-amidite 1917 (350 mg, 0.614
mmol) were separately dried by coevaporating (×3) with dry benzene.
They were then combined in freshly prepared dry MeCN (4 mL). To
this mixture was added 1H-tetrazole (61.5 mg, 0.88 mmol) at -30 °C
under an argon atmosphere. After 5 min, the ice bath was removed.
The mixture was further stirred for 2 h at room temperature, and then
the mixture was diluted with ethyl acetate. The organic phase was
washed with saturated aqueous NaHCO3 (×2), dried with MgSO4, and
evaporated. This phosphite derivative was used in the next steps without
purification. To a solution of phosphite in dry MeCN (4 mL) was added
t-BuOOH (2.5 M in toluene, 1.2 mL, 3.0 mmol), and the mixture was
stirred at room temperature for 1 h. After addition of dimethyl sulfide
(220 µL, 3.0 mM), to the mixture was added 1,8-diazabicyclo-[5.4.0]-
7-undecene (40 µL, 0.27 mmol). The mixture was stirred at room
temperature. After 5 min, to this mixture were added NaOMe (190
mg, 3.5 mmol) and MeOH:H2O ) 1:2 (6 mL). After 20 h, the mixture
was washed with CH2Cl2 (×2), and the water layer was evaporated at
25 °C. Purification of the residue on a silica gel column chromatography
(n-PrOH:20 mM NH4OH ) 3:1) and a gel permeation column
chromatography (Sephadex G-15, 1 cm × 60 cm, 20 mM NH4OH)
gave CMP-[3′′- 13C]-NeuAc 20 (37 mg, 35%, 4 steps) as an amorphous
1
mass; H NMR (D2O, 50 mM ND4CO3) δ 8.04 (d, 1H, J5,6 7.6 Hz,
H-6), 6.20 (d, 1H, J5,6 7.6 Hz, H-5), 6.07 (d, 1H, J1′,2′ 4.5 Hz, H-1′),
4.44 ∼ 4.28 (m, 5H, H-2′, 3′. 4′, 5a′, 5b′), 4.22 (dd, 1H, J5′′,6′′ 10.4 Hz,
J6′′,7′′ 0.9 Hz, H-6′′), 4.20-4.06 (m, 1H, H-4′′), 4.05-3.96 (m, 1H, H-5′′,
H-8′′), 3.96 (dd, 1H, J9a′′,8′′ 2.5 Hz, Jgem 11.8 Hz, H-9a′′), 3.70 (dd, 1H,
J9b′′,8′′ 6.4 Hz, Jgem 11.8 Hz, H-9b′′), 3.53 (dd, 1H, J6′′,7′′ 0.9 Hz, J7′′,8′′
9.6 Hz, H-7′′), 2.57 (ddd, 1H, J3e′′,4′′ 4.7 Hz, Jgem 13.3 Hz, J3e′′,C3′′ 135.5
Hz, H-3e′′), 2.13 (s, 3H, NHAc), 1.72 (dddd, 1H, J3a′′,4′′ 12.1 Hz, J3a′′,P
13
5.7 Hz, Jgem 13.3 Hz, J3a′′,C3′′ 128.0 Hz, H-3a′′); HRMS calcd for C19
-
CH29N4Na2O16P (M + Na+) 682.1044, found 682.1005.
[U-13C]-Gal-â-x-Ovalbumin (22). The transfer reaction was per-
formed in a solution of HEPES buffer (100 mM, pH 7.0, 1.0 mL)
containing UDP-[U-13C]-Glc18 (6.0 mg, 10.1 µmol), ovalbumin 21 (153
mg, 3.4 nmol), UDP-glucose-4-epimerase (200 mU), boVine â-1,4-
galactosyltransferase (100 mU), MnCl2 (50 mM), CIAP (50 U), and
the mixture was incubated at 37 °C for 24 h. The mixture was then
passed through Molcut-L filter (Millipore, 10 000 cut). The residue in
the filter was further washed with water (2 mL × 5) and then analyzed
by NMR. The NMR data are summarized in Table 3.
[3-13C]-NeuAc-r-(2f3)-[U-13C]-Gal-â-x-Ovalbumin (23). The
transfer reaction was performed in a solution of sodium cacodylate
buffer (50 mM, pH 6.0, 1.0 mL) containing CMP-[3′′-13C]-NeuAc 20
(2.1 mg, 3.4 µmol), [U-13C]-Gal-ovalbumin 22 (46.5 mg), CIAP (125U),
rat recombinant R-2,3-(N)-sialyltransferase (50 mU), and the mixture
was incubated at 37 °C for 16 h. The mixture was passed through
Molcut-L filter (UFP1LGCBK, Millipore Ltd., 10 000 cut). This transfer
reaction was repeated for more two times. The 50% of galactoside was
estimated to be sialylated by HMQC spectrum. To this mixture of asialo-
and sialoglycoprotein in a solution of sodium cacodylate buffer (50
mM, 500 µL, pH 6.0) was added NaN3 (1.0 mg) and D. pneumoniae
â-galactosidase (50 mU). The mixture was incubated at 30 °C for 24
h. The mixture was passed through Molcut-L filter (Millipore, 10 000
cut). The residue in the filter was washed with water (500 µL × 5)
and then analyzed by NMR. The NMR data are summarized in Table
3.
(30) Sabesan, S.; Paulson, J. C. J. Am. Chem. Soc. 1986, 108, 2068-
2080.
(31) (a) Geen, H.; Freeman, R. J. Magn. Reson. 1995, 93, 93-141. (b)
Roumestand, C.; Mispelter, J.; Austruy, C.; Canet, D. J. Magn. Reson.,
Ser. B 1995, 109, 153-163.
[U-13C]-Gal-â-(1f4)-GlcNAc-â-OMe (25). The transfer reaction
was performed in a solution of HEPES buffer (100 mM, pH 7.0, 500

5694 J. Am. Chem. Soc., Vol. 122, No. 24, 2000
Miyazaki et al.
Figure 13. Pulse sequences of modified 1D HMQC, HMQC-NOESY, 1D HSQC-TOCSY, 1D HSQC-TOCSY-NOESY, and 1D HSQC-TOCSY-
NOESY-TOCSY.
(1.05 ms), d10 (mixing time); φ1 ) x, φ2 ) x, φ3 ) y, φ4 ) 2x,
2(-x), φ5 ) x, φ6 ) x, -x, φ7 ) y, -y, φ8 ) x, φ9 ) -y, φ10 ) x,
φ11 ) 4x, 4(x), φ12 ) x, φ13 ) x, φ14 ) y, φ15 ) x, φ26 ) y, receiver
phase ) x, -x, x, -x, -x, x, -x, x. The following gradient ration were
calculated and then used: gz1:gz2:gz3:gz4 ) 40:15:10.05:-55. (E) 1D
HSQC-TOCSY-NOESY-TOCSY experiment. Dulations for RE-BURP
(90°) for H-7NeuAc, and E-BURP (90°) for C-3NeuAc selective pulses31
were 50 ms and 4 ms, respectively. Both spin locking (MLEV17) times
were 50 ms.32 The following delay time and phase cycling were
applied: d1 (1 s), d2 (1.6 ms), d3 (3 µs), d4 (1.1 ms), d5 (3 µs), d6
(100 µs), d7 (26 µs), d8 (3 µs), d9 (1.05 ms), d10 (mixing time), d11
(104 µs); φ1 ) x, φ2 ) x, φ3 ) y, φ4 ) 2x, 2(-x), φ5 ) x, φ6 ) x,
-x, φ7 ) y, -y, φ8 ) x, φ9 ) -y, φ10 ) x, φ11 ) 4x, 4(x), φ12 )
x, φ13 ) x, φ14 ) y, φ15 ) x, φ26 ) y, receiver phase ) x, -x, x,
-x, -x, x, -x, x. The followin gradient ratio were calculated and then
used: gz1:gz2:gz3:gz4:gz5:gz6 ) 40:15:10.05: -55:32:22. For 2D
HSQC-TOCSY-NOESY-TOCSY, the last TOCSY was developped to
the second dimension by TPPI mode.
AMBER*34 which was modified to contain substructure parameters for
sialic acid according to A. Bernardi et al.15k Conformational analysis
of NeuAc-R-(2f3)-Gal-â-(1f4)-GlcNAc-â- sequence was performed
in GB/SA35 water using the Monte Carlo procedure. In this conforma-
tional search, aglycon at the reducing end was used methyl glycoside.
We searched the lowest-energy conformation satisfying three distance
derived by NOE experiments (H-3axNeuAc/H-3Gal ) 2.7 ( 0.0 Å, H-1Gal
/
H-6GlcNAc ) 3.0 ( 0.3 Å, H-1Gal/H-6′GlcNAc ) 3.0 ( 0.3 Å). These
distance constraints were applied with the force constant of 250 kJ/
(mol‚Å2). The Monte Carlo simulation was set to vary all dihedral
angles except H6NeuAc-C6NeuAc-C7NeuAc-H7NeuAc (-72°), H7NeuAc
-
C7NeuAc-C8NeuAc-H8NeuAc (-158°), H2GlcNAc-C2GlcNAc-N2GlcNAc-NH-
(180°), and H5NeuAc-C5NeuAc-N5NeuAc-NHNeuAc (180°). With
GlcNAc
default force constants, these four dihedral angles estimated by coupling
constants (Table 3) were constrained. The Monte Carlo search was
conducted with a total 19 000 search steps using TNCG minimization
to gradient convergence (<0.05 kJ/mol). A total 2276 unique confor-
mations were saved. The lowest-energy conformer was found three
times and there were 14 additional conformaers within 2.0 kcal/mol.
The lowest-energy conformer is shown in Figure 12.
13C Spin-Lattice Relaxation Time (T1). T1 measurements were
obsd
carried out by using the inversion-recovery technique and the T1
values were determined from sixteen delays (τ ) 0.10, 0.15, 0.20, 0.25,
0.30, 0.35, 0.40, 0.45, 0.50, 0.55, 0.60, 0.65, 0.70, 0.80, 0.90, and 1.00
s) and a fully relaxed spectrum (with delay time between 180° and 90°
pulses ) T1 × 5 (s), spectral width ) 260 ppm, pulse width (90°) )
13 µs, data size ) 32K points) using the fit equation of I[τ] ) A
exp(-τ/T1) (I[τ]: peak intensity of when delay was used τ s). NOEs (η
+ 1) were measured from absolute integrated intensities using the gated
decoupling technique.24b T1DD values were calculated from the experi-
Acknowledgment. This work was partially supported by a
Grant-in-Aid for Encouragement of Young Scientists from the
Ministry of Education, Science, Sports and Culture, Japan (NO.
09780531) and the Takamura Foundation. The authors would
like to thank Mr. Masayoshi Kusama for mass spectroscopic
analysis.
obsd
DD
mentally determined T1
and η values using the equation T1
)
JA994211J
1.988 T1obsd/η.24
Computational Methods. Molecular mechanics calculations were
(33) Mohamadi, F.; Richards, N. G. J.; Guida, W. C.; Liskamp, R.;
Lipton, M.; Caufield, C.; Chang, G.; Hendrickson, T.; Still, W. C. J. Comput.
Chem. 1990, 11, 440-467.
(34) Homans, S. W. Biochemistry 1990, 29, 9110-9118.
(35) Still, W. C.; Tempczyk, A.; Hawley, R. C.; Hendrickson, T. J. Am.
Chem. Soc. 1990, 112, 6127-6129.
performed using MacroModel 5.5,33 and the force field used was
(32) (a) Stonehouse, J.; Adell, P.; Keeler, J.; Shaka, A. J. J. Am. Chem.
Soc. 1994, 116, 6037-6038. (b) Stott, K.; Stonehouse, J.; Keeler, J.; Hwang,
T.-L.; Shaka, A. J. J. Am. Chem. Soc. 1995, 117, 4199-4200.